CN117065176A - Rodent sleep deprivation device and deprivation system based on machine vision - Google Patents
Rodent sleep deprivation device and deprivation system based on machine vision Download PDFInfo
- Publication number
- CN117065176A CN117065176A CN202311328800.1A CN202311328800A CN117065176A CN 117065176 A CN117065176 A CN 117065176A CN 202311328800 A CN202311328800 A CN 202311328800A CN 117065176 A CN117065176 A CN 117065176A
- Authority
- CN
- China
- Prior art keywords
- sleep
- deprivation
- animal
- sleep deprivation
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 208000010340 Sleep Deprivation Diseases 0.000 title claims abstract description 164
- 241000283984 Rodentia Species 0.000 title claims abstract description 42
- 230000007958 sleep Effects 0.000 claims abstract description 135
- 238000010171 animal model Methods 0.000 claims abstract description 133
- 241001465754 Metazoa Species 0.000 claims abstract description 101
- 238000012545 processing Methods 0.000 claims abstract description 49
- 238000004458 analytical method Methods 0.000 claims abstract description 27
- 230000008859 change Effects 0.000 claims abstract description 24
- 238000010191 image analysis Methods 0.000 claims description 16
- 238000013528 artificial neural network Methods 0.000 claims description 12
- 238000013135 deep learning Methods 0.000 claims description 9
- 206010062519 Poor quality sleep Diseases 0.000 claims description 8
- 230000036760 body temperature Effects 0.000 claims description 7
- 238000009825 accumulation Methods 0.000 claims description 6
- 230000005540 biological transmission Effects 0.000 claims description 2
- 241000699670 Mus sp. Species 0.000 description 29
- 238000000034 method Methods 0.000 description 23
- 241000699666 Mus <mouse, genus> Species 0.000 description 22
- 238000002474 experimental method Methods 0.000 description 16
- 230000000875 corresponding effect Effects 0.000 description 15
- 230000008569 process Effects 0.000 description 14
- 230000006399 behavior Effects 0.000 description 6
- 238000012549 training Methods 0.000 description 6
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 6
- 230000001276 controlling effect Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 230000033001 locomotion Effects 0.000 description 5
- 238000012360 testing method Methods 0.000 description 5
- 230000003542 behavioural effect Effects 0.000 description 4
- 210000004556 brain Anatomy 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000003370 grooming effect Effects 0.000 description 4
- 238000001514 detection method Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 238000005192 partition Methods 0.000 description 3
- 208000019116 sleep disease Diseases 0.000 description 3
- QGZKDVFQNNGYKY-UHFFFAOYSA-N Ammonia Chemical compound N QGZKDVFQNNGYKY-UHFFFAOYSA-N 0.000 description 2
- 230000037007 arousal Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000033228 biological regulation Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000010924 continuous production Methods 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 230000003111 delayed effect Effects 0.000 description 2
- 239000013013 elastic material Substances 0.000 description 2
- 230000004907 flux Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 208000015181 infectious disease Diseases 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000011176 pooling Methods 0.000 description 2
- 230000000384 rearing effect Effects 0.000 description 2
- 230000001105 regulatory effect Effects 0.000 description 2
- 230000008844 regulatory mechanism Effects 0.000 description 2
- 230000003938 response to stress Effects 0.000 description 2
- 230000033764 rhythmic process Effects 0.000 description 2
- 230000004617 sleep duration Effects 0.000 description 2
- 208000020685 sleep-wake disease Diseases 0.000 description 2
- 229910001220 stainless steel Inorganic materials 0.000 description 2
- 239000010935 stainless steel Substances 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 102000008867 ARNTL Transcription Factors Human genes 0.000 description 1
- 108010088547 ARNTL Transcription Factors Proteins 0.000 description 1
- 208000024827 Alzheimer disease Diseases 0.000 description 1
- 208000019901 Anxiety disease Diseases 0.000 description 1
- 206010003805 Autism Diseases 0.000 description 1
- 208000020706 Autistic disease Diseases 0.000 description 1
- 101150038243 CLOCK gene Proteins 0.000 description 1
- 102000016267 Leptin Human genes 0.000 description 1
- 108010092277 Leptin Proteins 0.000 description 1
- 239000004677 Nylon Substances 0.000 description 1
- 241000700159 Rattus Species 0.000 description 1
- 208000013738 Sleep Initiation and Maintenance disease Diseases 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 229910021529 ammonia Inorganic materials 0.000 description 1
- 230000036506 anxiety Effects 0.000 description 1
- 230000006400 anxiety behaviour Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000004397 blinking Effects 0.000 description 1
- 230000019771 cognition Effects 0.000 description 1
- 230000003930 cognitive ability Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000003651 drinking water Substances 0.000 description 1
- 235000020188 drinking water Nutrition 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000008451 emotion Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000036039 immunity Effects 0.000 description 1
- 206010022437 insomnia Diseases 0.000 description 1
- NRYBAZVQPHGZNS-ZSOCWYAHSA-N leptin Chemical compound O=C([C@H](CO)NC(=O)[C@H](CC(C)C)NC(=O)[C@H](CC(O)=O)NC(=O)[C@H](CC(C)C)NC(=O)[C@H](CCC(N)=O)NC(=O)[C@H](CC=1C2=CC=CC=C2NC=1)NC(=O)[C@H](CC(C)C)NC(=O)[C@@H](NC(=O)[C@H](CC(O)=O)NC(=O)[C@H](CCC(N)=O)NC(=O)[C@H](CC(C)C)NC(=O)[C@H](CO)NC(=O)CNC(=O)[C@H](CCC(N)=O)NC(=O)[C@@H](N)CC(C)C)CCSC)N1CCC[C@H]1C(=O)NCC(=O)N[C@@H](CS)C(O)=O NRYBAZVQPHGZNS-ZSOCWYAHSA-N 0.000 description 1
- 229940039781 leptin Drugs 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 229920001778 nylon Polymers 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008560 physiological behavior Effects 0.000 description 1
- 230000035790 physiological processes and functions Effects 0.000 description 1
- 238000004886 process control Methods 0.000 description 1
- 208000020016 psychiatric disease Diseases 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 208000022925 sleep disturbance Diseases 0.000 description 1
- 230000004622 sleep time Effects 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 238000010415 tidying Methods 0.000 description 1
- 238000013518 transcription Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01K—ANIMAL HUSBANDRY; CARE OF BIRDS, FISHES, INSECTS; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
- A01K1/00—Housing animals; Equipment therefor
- A01K1/02—Pigsties; Dog-kennels; Rabbit-hutches or the like
- A01K1/03—Housing for domestic or laboratory animals
- A01K1/031—Cages for laboratory animals; Cages for measuring metabolism of animals
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01K—ANIMAL HUSBANDRY; CARE OF BIRDS, FISHES, INSECTS; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
- A01K29/00—Other apparatus for animal husbandry
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01K—ANIMAL HUSBANDRY; CARE OF BIRDS, FISHES, INSECTS; FISHING; REARING OR BREEDING ANIMALS, NOT OTHERWISE PROVIDED FOR; NEW BREEDS OF ANIMALS
- A01K29/00—Other apparatus for animal husbandry
- A01K29/005—Monitoring or measuring activity, e.g. detecting heat or mating
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/0033—Features or image-related aspects of imaging apparatus classified in A61B5/00, e.g. for MRI, optical tomography or impedance tomography apparatus; arrangements of imaging apparatus in a room
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/48—Other medical applications
- A61B5/4806—Sleep evaluation
- A61B5/4809—Sleep detection, i.e. determining whether a subject is asleep or not
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61M—DEVICES FOR INTRODUCING MEDIA INTO, OR ONTO, THE BODY; DEVICES FOR TRANSDUCING BODY MEDIA OR FOR TAKING MEDIA FROM THE BODY; DEVICES FOR PRODUCING OR ENDING SLEEP OR STUPOR
- A61M21/00—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
- G06V10/245—Aligning, centring, orientation detection or correction of the image by locating a pattern; Special marks for positioning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/57—Mechanical or electrical details of cameras or camera modules specially adapted for being embedded in other devices
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/64—Computer-aided capture of images, e.g. transfer from script file into camera, check of taken image quality, advice or proposal for image composition or decision on when to take image
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/90—Arrangement of cameras or camera modules, e.g. multiple cameras in TV studios or sports stadiums
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B2503/00—Evaluating a particular growth phase or type of persons or animals
- A61B2503/40—Animals
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B2503/00—Evaluating a particular growth phase or type of persons or animals
- A61B2503/42—Evaluating a particular growth phase or type of persons or animals for laboratory research
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61M—DEVICES FOR INTRODUCING MEDIA INTO, OR ONTO, THE BODY; DEVICES FOR TRANSDUCING BODY MEDIA OR FOR TAKING MEDIA FROM THE BODY; DEVICES FOR PRODUCING OR ENDING SLEEP OR STUPOR
- A61M21/00—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis
- A61M2021/0005—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis by the use of a particular sense, or stimulus
- A61M2021/0022—Other devices or methods to cause a change in the state of consciousness; Devices for producing or ending sleep by mechanical, optical, or acoustical means, e.g. for hypnosis by the use of a particular sense, or stimulus by the tactile sense, e.g. vibrations
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61M—DEVICES FOR INTRODUCING MEDIA INTO, OR ONTO, THE BODY; DEVICES FOR TRANSDUCING BODY MEDIA OR FOR TAKING MEDIA FROM THE BODY; DEVICES FOR PRODUCING OR ENDING SLEEP OR STUPOR
- A61M2205/00—General characteristics of the apparatus
- A61M2205/10—General characteristics of the apparatus with powered movement mechanisms
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61M—DEVICES FOR INTRODUCING MEDIA INTO, OR ONTO, THE BODY; DEVICES FOR TRANSDUCING BODY MEDIA OR FOR TAKING MEDIA FROM THE BODY; DEVICES FOR PRODUCING OR ENDING SLEEP OR STUPOR
- A61M2250/00—Specially adapted for animals
Abstract
The invention discloses a rodent sleep deprivation device and a deprivation system based on machine vision, wherein the deprivation device comprises a cage box, an image collector, image processing analysis equipment and a mechanical arm, wherein the image collector is used for collecting images of experimental animals in a preset period according to a time sequence to simulate human observation of experimental staff, the image processing analysis equipment is used for determining sleep states of the experimental animals and positioning coordinates corresponding to the experimental animals according to the change of extracted image features along with time, generating deprivation instructions and positioning coordinates and transmitting the deprivation instructions and the positioning coordinates to the mechanical arm, and the mechanical arm is used for carrying out sleep deprivation on the sleeping animals with the positioning coordinates according to a preset track after receiving the deprivation instructions and the positioning coordinates, so that the sleeping animals can be identified and accurately positioned and disturbed under the living environment condition without influencing the normal living environment of the experimental animals based on the machine vision, and meanwhile, high-throughput deprivation can be realized.
Description
Technical Field
The invention belongs to the field of animal behavioural experimental devices, and particularly relates to a rodent sleep deprivation device and a rodent sleep deprivation system based on machine vision.
Background
The chinese sleep study report 2023 issued in the past shows that the average sleep duration per night for the 2022 chinese adult is 7.4 hours, with the average sleep duration per night for nearly half of the population (47.55%) being less than 8 hours. In recent years, sleep disorder patients show an increasing trend, and sleep is very important for maintaining normal physiological functions of the brain, and lack of sleep can lead to significant decline of cognitive ability, immunity and emotion control ability of people; and many serious mental diseases such as Alzheimer's disease, autism, depression and the like are accompanied by insomnia symptoms.
The current mainstream view considers that sleep is regulated by a rhythm-steady state double-process model, namely, the rhythm process regulates the occurrence time of sleep, and the steady state process controls the quality and quantity of sleep. The rhythm regulation process is mainly controlled by negative feedback of core biological CLOCK gene translocation-transcription such as PER/CRY, BMAL1/CLOCK, etc., but the current cognition is limited for steady-state regulation process. In order to analyze the sleep steady-state regulation mechanism, the most basic and direct means is a sleep deprivation experiment, and the current sleep deprivation experiment is mainly divided into horizontal stage sleep deprivation, mechanical device deprivation, artificial deprivation and other modes.
The experimental mice are commonly used experimental animals, and the padding in the rearing cages can absorb excrement of the experimental mice in the rearing process of the mice, maintain the sanitation and hygiene of the cage boxes and the experimental mice, reduce ammonia concentration, maintain proper temperature and keep the interior of the cage boxes dry and comfortable, thereby being one of important factors affecting animal health and animal experimental results. The horizontal stage sleep deprivation utilizes the characteristic of the water of the mice to carry out sleep deprivation, is easy to generate stress reaction, can influence the scientificity of the sleep deprivation experiment and can change the normal living environment of the mice in the experimental process. Mechanical device deprives a device for testing rodent sleep disturbance and record as in patent CN 115336538A, the device in the patent converts the vibration signal of the test animal into an electric signal through a pressure sensor, judges whether the test animal is sleeping or moving through detecting the magnitude of the electric signal, and performs mechanical disturbance if the test animal is in a sleeping state, but the vibration signal detected by the device is easily influenced by pad vibration, and is easy to cause misjudgment, so the device is not suitable for pad, and the normal living environment of mice is influenced; the device not only interferes with sleeping mice, but also interferes with active mice, and stress response is easy to generate.
Although the artificial deprivation can not influence the normal life of the experimental animal, the artificial deprivation depends on the artificial observation to judge that the experimental animal enters a sleep state to carry out interference deprivation, so that an experimenter is required to observe the behavior of the experimental animal all the time, the sleep deprivation effect is easy to be generated for the experimenter, and a plurality of experimental animals are difficult to be simultaneously considered, so that the artificial deprivation is generally suitable for short-term sleep deprivation experiments. Therefore, the device capable of automatically depriving sleep on the premise of not affecting the normal living environment of experimental animals is researched, and is beneficial to research of sleep steady-state regulation mechanism and research and development of drugs for improving sleep disorder.
Disclosure of Invention
In order to overcome the above drawbacks or improvements of the prior art, the present invention provides a rodent sleep deprivation device and a rodent sleep deprivation system based on machine vision, which aims to provide a rodent sleep deprivation device based on machine vision, wherein the deprivation device is based on machine vision to simulate artificial observation to avoid the influence of the existing intelligent deprivation device on the normal life of mice, thereby solving the technical problem that the existing intelligent judgment method can influence the normal physiological behavior of the mice.
To achieve the above object, according to one aspect of the present invention, there is provided a machine vision-based rodent sleep deprivation device including a cage, an image collector, an image processing analysis device and a robot arm;
The image collector is arranged at the periphery of the cage box and is used for collecting images of a plurality of experimental animals according to a time sequence in a collecting period and transmitting the images to the image processing and analyzing equipment;
the image processing analysis equipment is used for extracting image features of the experimental animals from the received multiple images, determining the sleep state of the experimental animals and positioning coordinates corresponding to the experimental animals according to the change of the extracted image features along with time, selecting one of the experimental animals in sleep as a deprived object in each time slot, and transmitting a sleep deprivation execution instruction and positioning coordinate information of the sleep animal to the mechanical arm; the image features comprise one or more of the body posture, eye state, tail morphology and animal position of the experimental animal;
the mechanical arm is used for receiving the sleep deprivation executing instruction and the positioning coordinate information transmitted by the image processing and analyzing equipment, and carrying out sleep deprivation on the sleeping animal at the positioning coordinate according to the sleep deprivation executing instruction.
Preferably, the rodent sleep deprivation device based on machine vision, wherein the image processing analysis device is based on neural network deep learning to combine the changes of the posture, the state of eyes, the tail shape and the animal position of the experimental animal to divide the state of the experimental animal into three kinds of sleep, wakefulness and uncertainty, and one animal in the sleep state is selected as the deprived object.
Preferably, the machine vision based rodent sleep deprivation device, wherein the image processing analysis device selects a deprived subject:
generating a quantitative index D according to the time length of the experimental animal entering the sleep state, wherein the specific formula is D=at 2 +bt+c, where a, b>0；
Sequentially selecting corresponding sleeping animals as deprived objects according to the size sequence of the D value, and clearing the D value when the sleeping animals wake up or complete deprivation once;
and selecting the sleeping animal as the deprived object when the D value accumulation amount of the sleeping animal is greater than or equal to a preset threshold.
Preferably, in the rodent sleep deprivation device based on machine vision, a deprivation component is arranged at one end of the mechanical arm, and a controller is arranged in the rodent sleep deprivation device and is used for receiving a sleep deprivation execution instruction and positioning coordinate information transmitted by the image processing and analyzing equipment, and controlling the mechanical arm to drive the deprivation component to deprive sleep of a sleeping animal at the positioning coordinate.
Preferably, the rodent sleep deprivation device based on machine vision, the image collector comprises a plurality of camera modules 4, the camera modules 4 comprise a network camera 401 and a supporting rod 404, wherein the upper end of the supporting rod 404 is connected with the network camera 401, and the lower end is connected with a cage.
Preferably, the rodent sleep deprivation device based on machine vision further comprises an infrared sensor for acquiring the body temperature of the experimental animal and transmitting to an image processing analysis device.
Preferably, the rodent sleep deprivation device based on machine vision, the deprivation component of which is also provided with a pressure sensor.
Preferably, the rodent sleep deprivation device based on machine vision is characterized in that a plurality of independent cage grooves are formed in the cage box.
According to another aspect of the present invention, there is provided a machine vision-based rodent sleep deprivation device deprivation system according to the present invention comprising an image acquisition module, an image analysis processing module and a control module;
the image acquisition module is used for acquiring images of a plurality of experimental animals acquired in the acquisition period and submitting the images to the image analysis processing module;
the image analysis processing module is used for extracting image characteristics of the experimental animal from the received multiple images, wherein the image characteristics comprise one or more of the body posture, the eye state, the tail shape and the animal position of the experimental animal;
determining the sleep state of the experimental animal and the positioning coordinates corresponding to the experimental animal according to the change of the extracted image features along with time, selecting one of the experimental animals in sleep as a deprived object in each time slot, generating a sleep deprivation executing instruction and the positioning coordinate information of the sleep animal, and transmitting the sleep deprivation executing instruction and the positioning coordinate information to a control module;
The control module is used for controlling the sleep deprivation device to carry out sleep deprivation on the experimental animal with the position of the positioning coordinate in sleep according to the received sleep deprivation execution instruction and the positioning coordinate information; the sleep deprivation device comprises a cage box and a mechanical arm.
Preferably, in the deprivation system of the rodent sleep deprivation device based on machine vision, the image analysis processing module is used for deep learning based on a neural network, combining the body posture, the eye state, the tail shape and the animal position change of the experimental animal to divide the state of the experimental animal into three types of sleep, wakefulness and uncertainty, selecting one animal in the sleep state as a deprived object, generating a sleep deprivation execution instruction and positioning coordinate information of the sleeping animal, and transmitting the sleep deprivation execution instruction and the positioning coordinate information to the control module.
Preferably, the image analysis processing module generates a quantitative index D according to the time period of the experimental animal entering the sleep state, and the specific formula is d=at 2 +bt+c, where a, b>0；
Sequentially selecting corresponding sleeping animals as deprived objects according to the size sequence of the D value, and clearing the D value when the sleeping animals wake up or complete deprivation once;
And selecting the sleeping animal as the deprived object when the D value accumulation amount of the sleeping animal is greater than or equal to a preset threshold.
Preferably, in the deprivation system of the rodent sleep deprivation device based on machine vision, a plurality of independent cage grooves are arranged in the cage box, and the mechanical arm performs sleep deprivation according to the following track:
submerging from the side wall of the cage groove where the deprived object is positioned to the bottom surface, rotating to the other side wall of the cage groove around the shaft of the mechanical arm, lifting to the top of the cage groove, resetting, completing one sleep deprivation, and waiting for executing instructions next time;
or submerging from the middle of the cage groove where the deprived object is positioned to the bottom surface, moving to the side wall of the cage groove once, moving to the other side, depriving reciprocally once, lifting to the top of the cage groove, resetting, completing sleep deprivation once, and waiting for executing instructions next time.
In general, the above technical solutions conceived by the present invention, compared with the prior art, enable the following beneficial effects to be obtained:
according to the rodent sleep deprivation device based on machine vision, as the image collector collects a plurality of images of the experimental animal in a period according to a time sequence, long-term artificial observation of an experimenter can be simulated, normal life of the mouse is not affected, and the sleep state of the experimental animal is determined according to the change of one or more parameters of the body posture, the eye state, the tail shape and the animal position of the experimental animal in the collected images along with the collection time, so that the animal can be prevented from being in sleep when being in licking with the body but the position is unchanged; particularly, the experimental animals are divided into three states of sleep, wakefulness and uncertainty, so that the judgment accuracy of the sleeping animals can be improved. In addition, the mechanical arm is adopted to interfere sleeping animals, so that personalized positioning accurate interference and high-flux interference can be realized, and stress response to other experimental animals is avoided.
Furthermore, the deprivation system provided by the invention adopts neural network deep learning, and judges whether the experimental animal is in a sleep state by combining the posture of the body, the state of eyes, the shape of tail and the change of one or more parameters in the animal position of the experimental animal.
Drawings
Fig. 1 is a schematic diagram of a machine vision based rodent sleep deprivation device of example 1;
in the figure, a 1-mechanical arm, a 2-cage wall, a 3-bottom plate and a 4-camera module are shown, wherein a 201-cage groove, a 202-water groove, a 401-network camera, a 402-support rod, a 403-vertical hinged fixing seat, a 404-support rod and a 405-fixing seat are shown;
FIG. 2 is a schematic diagram of a deprivation component;
101-flange, 102-strut, 103-tee, 104-sleep deprivation rod;
fig. 3 is a schematic diagram of a machine vision based rodent sleep deprivation system.
Detailed Description
The present invention will be described in further detail with reference to the following examples in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention. In addition, the technical features of the embodiments of the present invention described below may be combined with each other as long as they do not collide with each other.
The invention provides a rodent sleep deprivation device based on machine vision, which comprises a cage box, an image collector, image processing analysis equipment and a mechanical arm, wherein the cage box is arranged on the cage box;
the image collector is provided with a connector connected with the switch, is arranged around the outer periphery of the cage box and is used for collecting images of a plurality of experimental animals in a period according to a time sequence and submitting the collected images to the image processing analysis equipment through the switch;
the image processing analysis equipment is used for extracting image features of the experimental animals from the received multiple images, determining the sleep state of the experimental animals and positioning coordinates corresponding to the experimental animals according to the change of the extracted image features along with time, selecting one of the experimental animals in sleep as a deprived object in each time slot, and transmitting a sleep deprivation execution instruction and positioning coordinate information of the sleep animal to the mechanical arm; the image features comprise one or more of the body posture, eye state, tail morphology and animal position of the experimental animal;
sleep is a continuous process, and the current sleep state judgment is generally divided into two states of wakefulness and sleep, and at the moment of judgment, the sleep state judgment is performed by detecting a certain parameter of the experimental animal and sleep activity. However, it is difficult to extract these parameters, such as brain electrical parameters, pressure parameters, without affecting the normal life of the experimental animal.
By adopting machine vision, the image data is extracted for judgment, so that the infection to the normal life of the experimental animal can be effectively avoided. Although the conventional machine vision means can capture relevant information to judge whether sleep deprivation is needed or not under the condition that normal life of an experimental animal is not affected, the conventional machine vision means only detects and judges a sleep state according to movement of a mouse position, however, due to continuous sleeping process, image acquisition and judgment are carried out at detection time, and accuracy is inevitably reduced. For example, it is easy to misjudge that an experimental animal is in a behavior state such as Grooming (Grooming) or Body Licking (Body Licking) to be in a sleep state, and even if the mechanical arm is not judged to be blocked, it is generally necessary to judge in a deprivation gap, and the result of the sleep deprivation experiment is easily affected.
The invention extracts the change of animal state along with time by adopting image data in a period of time based on machine vision, thereby reflecting the whole sleeping process, utilizing neural network deep learning to judge the sleeping state, and solving the problem of low accuracy of the machine vision when judging the sleeping state by using static images. Meanwhile, according to the sleep progress characteristics, the experimental animal is divided into three states of sleep, wake-up and uncertainty, but not the sleep/wake-up states when the electroencephalogram judgment is adopted, so that the method is more in line with the sleep process of the animal reflected in the behavioural aspect, and a more accurate sleep state judgment result can be obtained.
After the experimental animal images are acquired, the sleep state of the experimental animal is determined by extracting one or more parameters of the body posture, the eye state, the tail shape and the animal position of the experimental animal along with the change of the acquisition time, and particularly, the sleep state of the experimental animal is combined with various parameters, so that the behaviors of finely distinguishing the sleep and the wake of the mice can be realized under the condition that the normal life of the experimental animal is not influenced, the misjudgment is reduced, the judgment accuracy of the sleep deprivation device on the sleeping mice is further improved, the personalized accurate interference can be realized, and the sleep deprivation can be carried out with high flux.
Preferably, the experimental animals are classified into sleep, wakefulness and uncertainty, one animal is selected from the sleep animals as a deprived object, and a sleep deprivation execution instruction and position and location information are generated and transmitted to the mechanical arm;
sleep is a continuous process, and the current sleep state judgment is generally divided into two states of wakefulness and sleep, and at the moment of judgment, the sleep state judgment is performed by detecting a certain parameter of the experimental animal and sleep activity. However, it is difficult to extract these parameters, such as brain electrical parameters, pressure parameters, without affecting the normal life of the experimental animal.
By adopting machine vision, the image data is extracted for judgment, so that the infection to the normal life of the experimental animal can be effectively avoided. Although the conventional machine vision means can capture relevant information to judge whether sleep deprivation is needed or not under the condition that normal life of an experimental animal is not affected, the conventional machine vision means only detects and judges a sleep state according to movement of a mouse position, however, due to continuous sleeping process, image acquisition and judgment are carried out at detection time, and accuracy is inevitably reduced. For example, it is easy to misjudge that an experimental animal is in a behavior state such as Grooming (Grooming) or Body Licking (Body Licking) to be in a sleep state, and even if the mechanical arm is not judged to be blocked, it is generally necessary to judge in a deprivation gap, and the result of the sleep deprivation experiment is easily affected.
The invention extracts the change of animal state along with time by adopting image data in a period of time based on machine vision, thereby reflecting the whole sleeping process, utilizing neural network deep learning to judge the sleeping state, and solving the problem of low accuracy of the machine vision when judging the sleeping state by using static images. Meanwhile, according to the sleep progress characteristics, the experimental animal is divided into three states of sleep, wake-up and uncertainty, but not the sleep/wake-up states when the electroencephalogram judgment is adopted, so that the method is more in line with the sleep process of the animal reflected in the behavioural aspect, and a more accurate sleep state judgment result can be obtained.
After the experimental animal images are acquired, the sleep state of the experimental animal is determined by extracting one or more parameters of the body posture, the eye state, the tail shape and the animal position of the experimental animal along with the change of the acquisition time, and particularly, the sleep state of the experimental animal is combined with various parameters, so that the behaviors of finely distinguishing the sleep and the wake of the mice can be realized under the condition that the normal life of the experimental animal is not influenced, the misjudgment is reduced, the judgment accuracy of the sleep deprivation device on the sleeping mice is further improved, the personalized accurate interference can be realized, and the sleep deprivation can be carried out with high flux.
One of the sleeping animals was selected as a deprived subject, specifically as follows:
if a single animal is asleep in the period, selecting the single animal as a deprived object, and generating a sleep deprivation executing instruction;
if a plurality of animals are in sleep in a period, one sleeping animal is extracted to carry out sleep deprivation, and the method is as follows:
generating a quantitative index D according to the time length of the experimental animal entering the sleep state, wherein the specific formula is D=at 2 +bt+c, where a, b>0；
And sequentially selecting the corresponding sleeping animals as deprived objects according to the size sequence of the D value, and clearing the D value when the sleeping animals wake up or complete deprivation once. If the D value is taken as probability, extracting an experimental animal as a deprived object and generating a sleep deprivation instruction, and resetting the value;
Wherein the D value has an upper limit, and when the D value accumulation amount of a certain sleeping animal is larger than or equal to a preset threshold value (obtained according to experimental experience), the sleeping animal is selected as a deprived object, and the sleep deprivation is prevented from being delayed too much.
To prevent frequent deprivation of an animal, interfering with the experimental procedure or causing anxiety to occur, sleep deprivation is performed only once during a time interval even though it is still asleep.
The mechanical arm is used for receiving the sleep deprivation executing instruction and the positioning coordinate information transmitted by the image processing and analyzing equipment, and carrying out sleep deprivation on the sleeping animal at the positioning coordinate according to the sleep deprivation executing instruction.
In some embodiments, a connection port connected with the switch is arranged outside the mechanical arm, a controller is arranged in the mechanical arm, the controller is integrated on the mechanical arm base, the mechanical arm base and the mechanical arm base are integrated, the mechanical arm controller is connected with the image processing analysis equipment through the switch, and the controller is used for receiving sleep deprivation execution instructions of the image processing analysis equipment to control the mechanical arm to interfere with the experimental animal in the sleep state of the positioning coordinates and deprive the sleep of the experimental animal.
In some embodiments, the connection port is a network port TCP/IP, and the switch is responsible for data communication and allocation of each link component, and the image collector and the image processing analysis equipment such as a host computer and a mechanical arm controller perform data communication through the switch.
Further, one end of the mechanical arm is provided with a deprivation component for disturbing the received experimental animal in the sleep state of the positioning coordinates according to the preset motion trail so as to deprive the sleep of the experimental animal.
When the device is used, the image collector and the mechanical arm are respectively connected with the image processing analysis equipment through the switch, wherein the image processing analysis equipment comprises an upper computer, the image processing analysis equipment extracts image characteristics of an experimental animal according to an acquired image, the image characteristics comprise one or more of the body gesture, the eye state (such as information of opening and closing, blinking and the like), the tail state and the animal position change of the experimental animal, the state of the experimental animal is divided into three types of sleep, wake-up and uncertainty according to the change of the extracted image characteristics, if the experimental animal is judged to be in a sleep state, one animal is selected to perform sleep deprivation, a sleep deprivation interference instruction and position positioning information are generated, and the interference instruction and the positioning information are transmitted to the mechanical arm controller through the switch; the mechanical arm controller is used for receiving world coordinates S (positioning coordinates) and sleep deprivation executing instructions sent by the upper computer after image processing analysis, and controlling the mechanical arm to drive the deprivation component to carry out sleep deprivation on the sleeping animal according to a preset track of the mechanical arm.
In some embodiments, the image collector comprises a plurality of camera modules 4 which are uniformly distributed around the outer periphery of the cage box, and the collected image information is transmitted to image processing analysis equipment such as a host computer through a switch;
the camera module 4 comprises a network camera 401, a support rod 402, a vertical hinged fixing seat 403, a support rod 404 and a fixing seat 405, wherein the network camera 401 is connected with one end of the support rod 402, the other end of the support rod 402 is connected with the upper end of the support rod 404 through the vertical hinged fixing seat 403, and the lower end of the support rod 404 is connected with the bottom plate 3 of the cage through the fixing seat 405.
In some embodiments the robotic arm is centrally located in the customized cage, and the deprivation component at one end is a deprivation rod comprising flange 101, strut 102, sleep deprivation rod 104; one end of the strut 102 is detachably connected with the flange 101, the other end is connected with the sleep deprivation rod 104, such as through a tee joint 103, and a pressure sensor is arranged at the flange end of the strut (the component can also be integrated at the tail end of the last output shaft of the mechanical arm and is not marked in fig. 1); the sleep deprivation rod is made of elastic materials such as a tough nylon rod, so that the damage of the sleep deprivation rod to the mice can be avoided.
In some embodiments, the end of the last output shaft of the mechanical arm (flanges are generally used in pairs, and a flange plate correspondingly matched with the end of the mechanical arm is also arranged at the end of the mechanical arm) is provided with a pressure sensor along the axial direction, so that the strength of the sleep deprivation rod is regulated and controlled, and the damage to an animal model is avoided.
In some embodiments, after the mechanical arm receives the instruction, the mechanical arm starts from an initial position and moves to the position above the deprived sleep animal, the motion of a person stamp is simulated, a force sensor is arranged at the front end of the mechanical arm, when the contact force between the deprivation rod and the animal is larger than or equal to a preset threshold, the preset threshold is judged by simulating the contact force when the person deprives the animal to sleep according to a pre-experiment, or the contact force is determined according to the sleep deprivation state of the animal, the mechanical arm drives the deprivation rod to move away from the animal, and finally the mechanical arm returns to the initial point, so that the sleep deprivation of the animal in one period is completed. The image collector starts the collection and transmission of the image in the next period.
The shape of the cage box is not limited, and the cage box can be circular, fan-shaped and other shapes, in some embodiments, the cage box is circular, and comprises a cage wall and a bottom plate 3, a plurality of matte partition boards can be further arranged in the cage wall to form an independent cage groove 201, as shown in fig. 1, mutual interference among experimental animals can be avoided, the matte partition boards can not reflect images of the experimental animals, misjudgment can not be caused for image recognition, and the experimental animals capable of realizing high-flux accurate interference with sleeping can not affect other experimental animals.
In some embodiments, a plurality of independent cage grooves 201 are formed in the cage, and the mechanical arm performs sleep deprivation according to the following preset track:
Submerging from the side wall of the cage groove where the deprived object is positioned to the bottom surface, rotating to the other side wall of the cage groove around the shaft of the mechanical arm, lifting to the top of the cage groove, resetting, completing one sleep deprivation, and waiting for executing instructions next time;
or submerging from the middle of the cage groove where the deprived object is positioned to the bottom surface, moving to the side wall of the cage groove once, moving to the other side, depriving reciprocally once, lifting to the top of the cage groove, resetting, completing sleep deprivation once, and waiting for executing instructions next time.
A water tank 202 is also arranged outside the preferred cage wall and is used for providing a water source for experimental animals so as to ensure the normal drinking water requirement of the experimental animals; the bottom plate 3 is preferably a stainless steel plate and is used for fixing the mechanical arm 1 and the camera module 4.
Furthermore, the rodent sleep deprivation device is also provided with an infrared sensor for collecting body temperature information of experimental animals, the information is submitted to an image processing analysis device through an exchanger, and the image processing analysis device analyzes and judges whether the mice are in a sleep state or not through the body temperature and image characteristic information jointly acquired by the image processing analysis device. The infrared sensor includes a leptin or MLX90640.
In the sleep state, the body temperature of the animal is reduced, the body temperature of the animal is found out according to the measurement result of the infrared sensor on the temperature distribution in the cage box area, a trend curve is drawn, and the sleep state of the experimental animal is judged in combination with the characteristic change of other images. Since temperature is a slow variable, and it is observed that the sleep state of an animal sometimes changes rapidly (in the order of 10 seconds), body temperature detection can only be used as an auxiliary index to judge the sleep state of an experimental animal.
In addition, the invention also provides a machine vision-based rodent sleep deprivation device deprivation system, which comprises an image acquisition module, an image analysis processing module and a control module;
the image acquisition module is used for acquiring images of a plurality of experimental animals acquired in a preset period and submitting the images to the image analysis processing module; in some embodiments the preset period is 10-15s; the sleep deprivation device acquires images of a plurality of experimental animals in a time sequence within a preset period.
The image analysis processing module is used for extracting image characteristics of the experimental animal from the received images, determining the sleep state of the experimental animal and the positioning coordinates corresponding to the experimental animal according to the change of the extracted image characteristics along with time, selecting one of the experimental animals in sleep as a deprived object in each time slot, generating a sleep deprivation execution instruction and the positioning coordinate information of the sleep animal, and transmitting the sleep deprivation execution instruction and the positioning coordinate information to the control module; the image features comprise one or more of the body posture, eye state, tail morphology and animal position of the experimental animal;
the control module is used for controlling the sleep deprivation device to carry out sleep deprivation on the experimental animal with the position of the positioning coordinate in sleep according to the received sleep deprivation execution instruction and the positioning coordinate information; the sleep deprivation device comprises a cage box and a mechanical arm.
Preferably, the experimental animal is classified into three types of sleep, wakefulness and uncertainty according to the change of the extracted image characteristic data, and a quantitative index D is generated according to the time period of the experimental animal entering the sleep state, and the quantitative index D is specifically as follows:
D=at 2 +bt+c, where a, b>0, t is the sleep time of the experimental animal in the acquisition period;
in the formula, a and b need to consider factors such as the running speed of the mechanical arm, and the like, optimization is carried out in an actual test, t is the time of the experimental animal entering a sleep state, and the value of t does not exceed a judging period.
Taking the D value as probability to extract the deprived object, generating a sleep deprivation executing instruction and positioning coordinate information corresponding to the sleep animal, and submitting the sleep deprivation executing instruction and the positioning coordinate information to a control module, wherein the specific information is as follows:
when only one experimental animal is in a sleep state, selecting the experimental animal as a deprived object, and judging that the probability of performing sleep deprivation is 100%;
when a plurality of experimental animals are in a sleep state, taking the D value of the experimental animals in the sleep state as probability, extracting one experimental animal to execute sleep deprivation; wherein the probability of performing sleep deprivation is determined to be 100% by the experimental animal in the sleep state with the accumulated amount of the D value being equal to or greater than the preset threshold.
When only a single experimental animal enters a sleep state or the D value accumulation amount of a certain sleep animal is larger than or equal to a preset threshold value, selecting the sleep animal as a deprived object;
when a plurality of experimental animals are in a sleep state at the same time, the time of entering the sleep state can influence the sequence of executing sleep deprivation, corresponding sleep animals are sequentially selected as deprived objects according to the size sequence of the D value, and when the sleep animals wake up or complete deprivation once, the D value is cleared.
The classification of the experimental animal state is based on neural network deep learning for judgment, specifically, a training set is determined by a method of data acquired in advance and manual labeling (Sleep, unsleep and Invild three types), and then the training set is input into a machine learning network for training a network model. In some embodiments, the experimental animals are classified specifically as follows:
taking as input a sequence of downsampled gray images, the images having 8 convolution kernels of size 1 x 1 through the convolution layers S1, which are dedicated to integrating the time information of each channel; then, the two full connection layers are entered for classification after two convolution-pooling operations in sequence; the Softmax classifier classifies the input results into three categories: sleep, unsleep or Invild; where Sleep represents "Sleep", unsleep represents "awake", and invald represents "uncertain".
In some embodiments, one end of the mechanical arm is provided with a deprivation component, and the control module is used for receiving a sleep deprivation execution instruction and a positioning coordinate of the deprived object, controlling the mechanical arm to drive the deprivation component to carry out sleep deprivation on the sleeping animal at the positioning position, and returning to wait for outputting a next instruction after the mechanical arm finishes one sleep deprivation.
In some embodiments, a plurality of independent cage grooves are formed in the cage, and the mechanical arm preferably performs deprivation according to any one of the following modes:
(1) the side wall of the cage groove where the deprived object is located is submerged to the bottom surface, then the side wall of the inner cavity of the cage groove is rotated to the other side wall of the inner cavity of the cage groove around the shaft of the mechanical arm, then the side wall of the cage groove is lifted to the top of the cage groove and then reset, one sleep deprivation is completed, the next instruction execution is waited, and the experimental animal is prevented from sleeping at a certain corner.
(2) The device is characterized in that the device is submerged from the middle of a cage groove where a deprived object is located to the bottom surface, moves to the side wall of the cage groove once, then moves to the other side, is lifted to the top of the cage groove after being deprived once in a reciprocating manner, and is reset after being deprived once, so that sleep deprivation once is completed, and the next instruction execution is waited.
The following are examples:
example 1 machine vision based rodent sleep deprivation device
The sleep deprivation device in this embodiment, as shown in fig. 1, includes a cage, an image collector, an image analysis processing device (not shown in the figure), and a mechanical arm;
The cage box is circular and comprises a cage wall 2 and a bottom plate 3, a plurality of partition plates are arranged in the cage box to form a plurality of independent cage grooves 201, water grooves 202 are arranged at positions corresponding to the cage grooves, and the bottom plate is made of stainless steel plates.
The image collector is connected with image analysis processing equipment (upper computer) through the switch, and its is specific camera module 4, camera module 4 includes network camera 401, branch 402, perpendicular articulated fixing base 403, branch 404, fixing base 405, and wherein network camera 401 is connected with the one end of branch 402, and the other end of branch 402 passes through perpendicular articulated fixing base 403 and is connected with the upper end of branch 404, and the lower extreme of branch 404 passes through fixing base 405 to be connected with the bottom plate 3 of cage box, evenly distributed five groups camera modules around the cage box is outer, is convenient for observe all experimental animals in the cage box simultaneously.
The arm is established in the positive central point position in the cage, establishes the controller in it, and the controller is integrated in the arm base, and both are an integer, is connected with image analysis processing equipment through the switch, and specific arm one end is equipped with deprives the pole, as shown in fig. 2, deprives the pole and includes flange 101, branch 102, sleep deprives pole 104, and wherein branch 102 one end is dismantled with flange 101 and is connected, and the other end is connected through tee bend 103 with sleep deprivation pole 104, sleep deprivation pole 104 is the elastic material.
Example 2 machine vision based rodent sleep deprivation device sleep deprivation of mice
The sleep deprivation device in example 1 was used in this example to perform sleep deprivation on the experimental mice, specifically as follows:
the schematic diagram of the sleep deprivation device is shown in figure 3, the experimental mice are placed in a cage box for standby food and water, one mouse is placed in each cage groove, and the cage grooves are numbered 1-12; the upper computer sets the number of the cage grooves for experiments or automatically identifies the number of the cage grooves through a camera; and (3) setting the total duration and speed of deprivation after adjusting and optimizing according to the effect of the pre-experiment, and starting operation.
The execution speed of the mechanical arm, the number of experimental cage cabos and the deprivation time can be set before experiments so as to adapt to sleep deprivation experiments of rats and mice with different durations.
And starting the camera module to acquire images according to a time sequence, acquiring the activity condition of the mice in the cage box, wherein the acquisition period is 14s, each image interval is 1 second, and transmitting the overlooking information of the acquired mice with 12 cages to the upper computer through the switch. The upper computer processes the collected image characteristic information of each cage of mice through Python or programming of other development tools such as C++, C#, C, java, matlab and the like, and specifically:
And converting the pixel coordinates of the mass center of the mouse obtained by the network camera into world coordinates S (positioning coordinates) through an internal reference matrix of the camera and an external reference matrix of the camera, wherein the world coordinates are coordinates obtained by the mechanical arm through teaching.
Collecting images through a camera, comprehensively considering the position change, the posture change, the morphology (such as tail and eyes) and the like of the experimental animal, and judging whether the mouse is in a sleep state or not; and judging the object to be deprived of sleep in the period according to the sleep judging condition, outputting a corresponding sleep deprivation instruction and a positioning coordinate, and sending the corresponding sleep deprivation instruction and the mouse positioning coordinate to the mechanical arm controller through the switch TCP/IP by the upper computer.
The mechanical arm controller receives a command to control the mechanical arm to generate a corresponding action command, and the action command performs sleep deprivation on the positioning coordinate sleep mouse through a track which is shown in advance, wherein the interference track is specifically as follows:
after the sleep deprivation rod descends to the bottom surface from the side wall of the cage groove corresponding to the deprived object, the sleep deprivation rod rotates to the other side wall in the cage groove around the mechanical arm shaft and then is lifted to the top of the cage box, one sleep deprivation is completed, the next sleep deprivation instruction is waited, and the track is reversely executed after the next recognition.
Or specifically as follows:
the sleep deprivation rod descends to the bottom surface from the middle of the cage groove corresponding to the deprived object, moves to the side wall of the cage groove once, then moves to the other side, deprives reciprocally once, and returns to the position after being lifted to the cage top, so that the sleep deprivation once is completed, and the next instruction execution output is waited.
The rodent sleep deprivation device is also provided with a monitor, and is used for feeding back the motion information and the track of the mechanical arm to a user in real time, so that the rodent sleep deprivation device is convenient to debug and observe.
After the experiment is finished, we record the brain electricity of the animal model and observe the complementation condition of the mice. Experiments show that the device has high sleep deprivation efficiency, and the phenomenon of stress does not appear in the mice after the experiments.
Example 3 deep learning construction of rodent sleep deprivation devices based on machine vision deprivation System
(1) Data acquisition
Data acquisition and pretreatment: the traditional machine vision method is utilized, the arousal state of the mouse is judged by taking the change of the mouse coordinates as a standard, the data with and without change of the mouse coordinates are selected, and the data with change of the mouse coordinates can be directly marked as arousal; the data of the unchanged mouse coordinates are manually calibrated by a professional experimenter, so that the workload of personnel data marking can be reduced, the phenomenon of unbalanced data distribution (particularly high-value data which is helpful for supplementing behaviors such as hair tidying, body licking and the like and low sleep distinction degree) can be solved, and the correctness of a data set can be ensured.
(2) Network model construction
The design architecture based on the Convolutional Neural Network (CNN) -LeNet-5 is improved, a convolution layer of 1 multiplied by 16 is added at the forefront end of the LetNet-5, the time information of the video (image sequence) is integrated and extracted, and the sizes of all layers of the LetNet5 are adjusted to adapt to the image resolution of the behavioural camera.
The LetNet-5 is mainly used for realizing the identification of a single-channel image, and extracting certain characteristics of the image (generated by the inside of the neural network) through 2 convolution layers; the two full connection layers realize classification of data results, and the image features extracted in the embodiment comprise body gestures, eye states (such as open and close, blink and other information), tail morphology, mouse position fluctuation and the like.
The classification process of the experimental animal is to determine a training set by a method of data acquired in advance and manual labeling (Sleep, unsleep and Invild three types), and then input the training set into a machine learning network to train a network model, and the method is specific:
using a down-sampled sequence of gray scale images as inputs, with inputs H x W x 8,H and W being the image resolution, limited to the hardware configuration of the camera, the number of channels 8 representing the input is a time sequence of 8 gray scale images, each image being spaced 1 second apart. The input image has 8 convolution kernels of size 1 x 1 through the convolution layer S1, dedicated to integrating the temporal information of each channel. And then sequentially performing convolution-pooling operation twice, and entering two full-connection layers for classification. The Softmax classifier classifies the input results into three categories: sleep, unsleep and Invill, where Sleep represents "Sleep", unsleep represents "awake", and Invill represents "uncertain".
When at least one mouse is detected to be in a sleep state and the mechanical arm is in a position state, selecting the one sleep mouse as a deprived object, specifically:
the sleeping mouse generates a quantitative index D according to the time period of entering the sleeping state, wherein the formula D=at 2 +bt+c, where a, b>0, t is the period of time that the mouse goes to sleep;
when one experimental mouse is in a sleep state, selecting the experimental mouse as a deprived object, and judging that the probability of carrying out sleep deprivation is 100%;
when a plurality of experimental mice are in a sleep state, taking the D value of each of the experimental mice in the sleep state as probability, extracting one object which is used for executing sleep deprivation, extracting the centroid pixel coordinate and the camera coordinate of the sleep mice, converting the centroid pixel coordinate and the camera coordinate into world coordinate S, and outputting a sleep deprivation execution instruction and world coordinate S positioning information.
To ensure that the latency of sleep deprivation is within a reasonable range, the present embodiment may also employ the following stochastic algorithm to assign sleep deprivation operations:
(1) defining a nonlinear function D (t) that characterizes the urgency of sleep deprivation (D "(t) > 0) of the animal, accumulating from the time the animal enters sleep, accumulating faster the longer the value accumulates, and clearing the value if the animal wakes up or completes a sleep deprivation;
(2) Randomly extracting a mouse with D (t) as probability to perform sleep deprivation when a plurality of animals are in a sleep state at the same time;
(3) d (t) is set with an upper limit, when D (t) of an animal reaches a preset value (determined by experimenters according to experience), random extraction is skipped, and sleep deprivation is directly carried out on the animal, so that the sleep deprivation is prevented from being delayed too much;
(4) dead time: to prevent frequent deprivation of an animal from interfering with the progress of the experiment, or causing anxiety behavior to occur, sleep deprivation (even though it is still asleep) is performed only once during a time interval.
(3) Automatic sleep deprivation
After receiving the sleep deprivation executing instruction and the positioning coordinate information, the control module controls the mechanical arm to move according to a preset track to interfere and deprive the sleep of the sleeping mouse.
Although the features extracted by the deprivation system are obtained by learning through the neural network, the convolution kernel is selected in advance according to the features in the process of initializing the neural network, so that the training amount is reduced, and the efficiency of the neural network is improved.
It will be readily appreciated by those skilled in the art that the foregoing description is merely a preferred embodiment of the invention and is not intended to limit the invention, but any modifications, equivalents, improvements or alternatives falling within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (10)
1. The rodent sleep deprivation device based on machine vision is characterized by comprising a cage box, an image collector, an image processing analysis device and a mechanical arm;
the image collector is arranged at the periphery of the cage box and is used for collecting images of a plurality of experimental animals according to a time sequence in a collecting period and transmitting the images to the image processing and analyzing equipment;
the image processing analysis equipment is used for extracting image features of the experimental animals from the received multiple images, determining the sleep state of the experimental animals and positioning coordinates corresponding to the experimental animals according to the change of the extracted image features along with time, selecting one of the experimental animals in sleep as a deprived object in each time slot, and transmitting a sleep deprivation execution instruction and positioning coordinate information of the sleep animal to the mechanical arm; the image features comprise one or more of the body posture, eye state, tail morphology and animal position of the experimental animal;
the mechanical arm is used for receiving the sleep deprivation executing instruction and the positioning coordinate information transmitted by the image processing and analyzing equipment, and carrying out sleep deprivation on the sleeping animal at the positioning coordinate according to the sleep deprivation executing instruction.
2. The machine vision based rodent sleep deprivation device of claim 1 wherein the image processing analysis means, based on neural network deep learning, combines changes in the pose of the experimental animal's body, eye state, tail morphology and animal position to divide the experimental animal's state into three categories, sleep, awake and uncertain, and selects one of the animals in sleep as the deprived subject.
3. The machine vision based rodent sleep deprivation device of claim 2 wherein the image processing analysis device selects a deprived subject as follows:
generating a quantitative index D according to the time length of the experimental animal entering the sleep state, wherein the specific formula is D=at 2 +bt+c, where a, b>0；
Sequentially selecting corresponding sleeping animals as deprived objects according to the size sequence of the D value, and clearing the D value when the sleeping animals wake up or complete deprivation once;
and selecting the sleeping animal as the deprived object when the D value accumulation amount of the sleeping animal is greater than or equal to a preset threshold.
4. The rodent sleep deprivation device based on machine vision of claim 3 wherein a deprivation element is provided at one end of the robot arm and a controller is provided in the robot arm for receiving sleep deprivation execution instructions and positioning coordinate information transmitted by the image processing and analysis device, controlling the robot arm to drive the deprivation element to deprive sleeping animals at the positioning coordinates.
5. The rodent sleep deprivation device based on machine vision of claim 1, wherein the image collector comprises a plurality of camera modules (4), the camera modules (4) comprising a webcam (401), a strut (404), wherein the upper end of the strut (404) is connected with the webcam (401) and the lower end thereof is connected with a cage.
6. The machine vision based rodent sleep deprivation device of claim 5 further comprising an infrared sensor for acquiring the body temperature of an experimental animal and transmitting to an image processing analysis device.
7. The machine vision-based rodent sleep deprivation device of any one of claims 1-5 comprising an image acquisition module, an image analysis processing module and a control module;
the image acquisition module is used for acquiring images of a plurality of experimental animals acquired in the acquisition period and submitting the images to the image analysis processing module;
the image analysis processing module is used for extracting image characteristics of the experimental animal from the received multiple images, wherein the image characteristics comprise one or more of the body posture, the eye state, the tail shape and the animal position of the experimental animal;
determining the sleep state of the experimental animal and the positioning coordinates corresponding to the experimental animal according to the change of the extracted image features along with time, selecting one of the experimental animals in sleep as a deprived object in each time slot, generating a sleep deprivation executing instruction and the positioning coordinate information of the sleep animal, and transmitting the sleep deprivation executing instruction and the positioning coordinate information to a control module;
The control module is used for controlling the sleep deprivation device to carry out sleep deprivation on the experimental animal with the position of the positioning coordinate in sleep according to the received sleep deprivation execution instruction and the positioning coordinate information; the sleep deprivation device comprises a cage box and a mechanical arm.
8. The machine vision-based rodent sleep deprivation device deprivation system of claim 7 wherein the image analysis processing module, based on neural network deep learning, combines the changes in the body posture, eye state, tail morphology and animal position of the laboratory animal to divide the laboratory animal state into three categories of sleep, wakefulness and uncertainty, selects one of the animals in sleep as a deprived subject, and generates sleep deprivation execution instructions and positional coordinate information of the sleeping animal for transmission to the control module.
9. The machine vision-based rodent sleep deprivation system of claim 8 wherein the image analysis processing module generates a quantitative indicator D based on the length of time the experimental animal enters sleep state, the specific formula being d=at 2 +bt+c, where a, b>0；
Sequentially selecting corresponding sleeping animals as deprived objects according to the size sequence of the D value, and clearing the D value when the sleeping animals wake up or complete deprivation once;
And selecting the sleeping animal as the deprived object when the D value accumulation amount of the sleeping animal is greater than or equal to a preset threshold.
10. The machine vision based rodent sleep deprivation device of claim 7 wherein a plurality of independent cage slots are provided in the cage, the robotic arm performing sleep deprivation according to the following trajectory:
submerging from the side wall of the cage groove where the deprived object is positioned to the bottom surface, rotating to the other side wall of the cage groove around the shaft of the mechanical arm, lifting to the top of the cage groove, resetting, completing one sleep deprivation, and waiting for executing instructions next time;
or submerging from the middle of the cage groove where the deprived object is positioned to the bottom surface, moving to the side wall of the cage groove once, moving to the other side, depriving reciprocally once, lifting to the top of the cage groove, resetting, completing sleep deprivation once, and waiting for executing instructions next time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311328800.1A CN117065176B (en) | 2023-10-15 | 2023-10-15 | Rodent sleep deprivation device and deprivation system based on machine vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311328800.1A CN117065176B (en) | 2023-10-15 | 2023-10-15 | Rodent sleep deprivation device and deprivation system based on machine vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117065176A true CN117065176A (en) | 2023-11-17 |
| CN117065176B CN117065176B (en) | 2023-12-19 |
Family
ID=88706353
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311328800.1A Active CN117065176B (en) | 2023-10-15 | 2023-10-15 | Rodent sleep deprivation device and deprivation system based on machine vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117065176B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20070016637A (en) * | 2005-08-04 | 2007-02-08 | 고려대학교 산학협력단 | Versatile sleep behavior analyzer |
| US10292369B1 (en) * | 2015-06-30 | 2019-05-21 | Vium, Inc. | Non-contact detection of physiological characteristics of experimental animals |
| CN209732247U (en) * | 2019-01-23 | 2019-12-06 | 浙江中医药大学 | Big mouse sleep deprivation equipment |
| CN212491079U (en) * | 2020-04-02 | 2021-02-09 | 中国科学院上海有机化学研究所 | Experimental animals sleep disturbing device |
| CN113546272A (en) * | 2020-04-02 | 2021-10-26 | 中国科学院上海有机化学研究所 | Experimental animals sleep disturbing device |
| CN113648503A (en) * | 2021-07-14 | 2021-11-16 | 广州医科大学 | Sleep interference device for rats and mice and use method thereof |
| WO2023155551A1 (en) * | 2022-02-18 | 2023-08-24 | 深圳先进技术研究院 | Animal sleep deprivation device |
-
2023
- 2023-10-15 CN CN202311328800.1A patent/CN117065176B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20070016637A (en) * | 2005-08-04 | 2007-02-08 | 고려대학교 산학협력단 | Versatile sleep behavior analyzer |
| US10292369B1 (en) * | 2015-06-30 | 2019-05-21 | Vium, Inc. | Non-contact detection of physiological characteristics of experimental animals |
| CN209732247U (en) * | 2019-01-23 | 2019-12-06 | 浙江中医药大学 | Big mouse sleep deprivation equipment |
| CN212491079U (en) * | 2020-04-02 | 2021-02-09 | 中国科学院上海有机化学研究所 | Experimental animals sleep disturbing device |
| CN113546272A (en) * | 2020-04-02 | 2021-10-26 | 中国科学院上海有机化学研究所 | Experimental animals sleep disturbing device |
| CN113648503A (en) * | 2021-07-14 | 2021-11-16 | 广州医科大学 | Sleep interference device for rats and mice and use method thereof |
| WO2023155551A1 (en) * | 2022-02-18 | 2023-08-24 | 深圳先进技术研究院 | Animal sleep deprivation device |
Non-Patent Citations (1)
| Title |
|---|
| 郑秀娟: "大鼠运动行为的自动分析", 医药卫生科技辑, no. 09, pages 1 - 54 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117065176B (en) | 2023-12-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bao et al. | Artificial intelligence in animal farming: A systematic literature review | |
| Chen et al. | Classification of drinking and drinker-playing in pigs by a video-based deep learning method | |
| CN101214168B (en) | Animal behavior monitoring and analyzing system | |
| EP3900891B1 (en) | Pet amusement control apparatus of robot and mobile robot | |
| CN106489790B (en) | Birds visual cognition learning training system | |
| JP2004532475A (en) | Systems and methods for monitoring behavioral information engineering | |
| CN102282570A (en) | System and method for stereo-view multiple animal behavior characterization | |
| CN110135242B (en) | Emotion recognition device and method based on low-resolution infrared thermal imaging depth perception | |
| US20160120153A1 (en) | Animal test device | |
| CN111134033A (en) | Intelligent animal feeder and method and system thereof | |
| CN204500970U (en) | For studying the experimental provision of batch toy anxiety/frightened behavior reaction | |
| CN112580552A (en) | Method and device for analyzing behavior of rats | |
| Zin et al. | A general video surveillance framework for animal behavior analysis | |
| CN111197845A (en) | Deep learning-based control method and system for air conditioner operation mode | |
| CN111046808B (en) | Analysis method for pig raising drinking water and playing drinking water device by adopting residual convolution neural network and long-term and short-term memory classification group | |
| CN109847168A (en) | Wearable fatigue detecting and interfering system | |
| CN111857355A (en) | Reading state monitoring feedback system | |
| CN117065176B (en) | Rodent sleep deprivation device and deprivation system based on machine vision | |
| CN113456063B (en) | Artificial intelligence-based dead chicken disease detection system and detection method | |
| CN111178172A (en) | Laboratory mouse sniffing action recognition method, module and system | |
| CN110263027A (en) | A kind of experiment mice drinking-water monitoring method and system | |
| CN106264546A (en) | One chews behavior monitoring method and device | |
| Wu et al. | Research on cow rumination monitoring based on new activity sensor | |
| CN210247950U (en) | Intelligent detection device and intelligent detection system | |
| CN110427848B (en) | Mental analysis system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Zhang Lei Inventor after: Zhao Bing Inventor after: Sang Haojun Inventor before: Zhao Bing Inventor before: Sang Haojun Inventor before: Zhang Lei |