CN110464611A - A kind of digitlization amblyopia enhancing training device and system and its related algorithm - Google Patents
A kind of digitlization amblyopia enhancing training device and system and its related algorithm Download PDFInfo
- Publication number
- CN110464611A CN110464611A CN201910667755.XA CN201910667755A CN110464611A CN 110464611 A CN110464611 A CN 110464611A CN 201910667755 A CN201910667755 A CN 201910667755A CN 110464611 A CN110464611 A CN 110464611A
- Authority
- CN
- China
- Prior art keywords
- pupil
- bright spot
- image
- component
- flashing
- 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.)
- Pending
Links
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
- A61H5/00—Exercisers for the eyes
Landscapes
- Health & Medical Sciences (AREA)
- Ophthalmology & Optometry (AREA)
- Epidemiology (AREA)
- Pain & Pain Management (AREA)
- Physical Education & Sports Medicine (AREA)
- Rehabilitation Therapy (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Rehabilitation Tools (AREA)
Abstract
The invention discloses a kind of digitlization amblyopia enhancing training device and system and its related algorithm, device includes an eyeshade;One visual channel, the visual channel are arranged in the eyeshade;One pupil image component, the pupil illumination image-forming assembly is perpendicular to visual channel;One flashing bright spot component, the light-emitting surface of the flashing bright spot component is apart from the eyeshade 250mm;One reflecting mirror, reflecting mirror described in the light-transmissive that the flashing bright spot component issues enter pupil and stimulate macular region, and pupil is will be in real time by the pupil image monitor component by the post-stimulatory situation of change of bright spot;The result images of one display screen, the pupil image component are shown on the display screen;And a host, the host control the flashing bright spot component and the flashing bright spot component and the display screen.
Description
Technical field
The present invention relates to the communications field, in particular to a kind of digitlization amblyopia enhances training device and system and its related calculation
Method.
Background technique
According to the latest definition of ophthalmology branch, Chinese Medical Association in 2010, amblyopia is referred mainly to, not bright in eyeball, visual pathway
In the case of aobvious organic disease, most preferably corrects defects of vision and the functional disease for the eyesight value being consistent with the puberty is not achieved.Such as: 2
The twenty-twenty vision in year 0.5 or so, 4 years old 0.8 or so, 5 years old or more > 1.0.
Eyeball is without significant organic pathology, and simple eye or eyes correct defects of vision and are still not achieved 1.0, is known as amblyopia, wherein
Most of amblyopias are caused by astigmatism, myopia or long sight etc., if easily developed to if children myopia is not controlled effectively
High myopia generates a series of pathology and sexually revises, and enters after person in middle and old age that will to become uncurable disease blind.According to statistics, more than 30%
Old blind disease be by high myopia posterior scleral stretch and caused by the severe complications such as concurrent glaucoma, cataract.Child strabismus is weak
If depending in early stage treatment in time, will likely also not develop into low visual acuity or blind disease.
In real life, amblyopia influences huge to our brings, not only influences whether our vision, sees that thing becomes
To fuzzy, child also will affect school work, life and later choose a job.Today, mother's net encyclopaedia was just weak to have a talk about in detail to everybody
Depending on harm, so that everybody more preferably understands amblyopia.
[normal development for influencing eye]
Amblyopia is relatively common in life, has when young child because malnutritive, ametropia equal and lead to amblyopia.Amblyopia meeting
Cause and aggravate myopia, influences the normal development of eye.
[lacking stereoscopic vision]
Influence of the amblyopia for people is bigger, especially child, if obtaining amblyopia, parent is had to it timely
Treatment.Amblyopia will cause binocular visual function lowly and lack stereoscopic vision.
[can not judge orientation]
Amblyopia harm is much larger than myopia.Simple myopia is seen and is far obscured, and sees close clear.And children with amblyopia, see far obscuring,
It sees close also unclear, while because stereoscopic vision is fuzzy, is unable to the orientation and distance of accurate judgement object.
[multi-party function is impaired]
Amblyopia will lead to the multi-party impaired of eye vision function, such as adjusting force, pan power, trace ability, spatial discrimination,
The multinomial very important visual function such as contrast sensitivity and hand eye coordination ability is impaired.
Summary of the invention
The technical problem to be solved in the present invention is to provide a kind of digitlization amblyopia enhancing training device and systems and its related
Algorithm can play excitement more stronger than daily vision input signal by specific position, certain strength vision input signal
Vision.It obtains amblyopia eye in the short time and is easier perceived, properer spatial frequency stimulation.It is yellow to amblyopia eye to play
The more effective special training of spot region.
In order to solve the above-mentioned technical problem, the technical solution of the present invention is as follows:
A kind of digitlization amblyopia training device, suitable for testing the pupil of training human eye, comprising:
One eyeshade;
One visual channel, the visual channel are arranged in the eyeshade;
One pupil image component, the pupil illumination image-forming assembly is perpendicular to visual channel;
One flashing bright spot component, the light-emitting surface of the flashing bright spot component is apart from the eyeshade 250mm;
One reflecting mirror, it is described flashing bright spot component issue light-transmissive described in reflecting mirror enter pupil and stimulate macula lutea
Region, pupil is will be in real time by the pupil image monitor component by the post-stimulatory situation of change of bright spot;
The result images of one display screen, the pupil image component are shown on the display screen;And
One host, the host control the flashing bright spot component and the flashing bright spot component and the display screen.
It preferably, further include an adjusting knob, the adjusting knob controls the flashing bright spot component, to adjust bright spot
Flicker frequency and color.
The present invention also provides a kind of digitlization amblyopia training systems, comprising:
One flicker fusion frequency measures subsystem, and the flicker fusion frequency measuring system generates a visual imaging optical path;
One computer control analysis subsystem, the computer control analysis subsystem controls current background light source, described
Computer control analysis subsystem control connects the flicker fusion frequency measuring system;And
One digitlization amblyopia training device, comprising:
One eyeshade;
One visual channel, the visual channel are arranged in the eyeshade, and the visual imaging optical path is arranged in the view
Feel channel;
One pupil image component, the pupil illumination image-forming assembly is perpendicular to visual channel;
One flashing bright spot component, the light-emitting surface of the flashing bright spot component is apart from the eyeshade 250mm, the computer control
System analysis subsystem controls the flashing bright spot component;
One reflecting mirror, it is described flashing bright spot component issue light-transmissive described in reflecting mirror enter pupil and stimulate macula lutea
Region, pupil is will be in real time by the pupil image monitor component by the post-stimulatory situation of change of bright spot;
The result images of one display screen, the pupil image component are shown on the display screen;And
One host, the host control the flashing bright spot component and the flashing bright spot component and the display screen.
Preferably, the flicker fusion frequency measuring system includes a source stimulating device and a monitoring camera, wherein described
Source stimulating device is provided with multiple bright spot light sources and background light source and a LED, and the source stimulating device uses integrating sphere
One of means of even light, the even light of even light pipe or the even smooth means of abrasive segments realize the LED to the light of the bright spot light source
According to uniform.
Preferably, the LED is circulating type LED.
Preferably, the computer control analysis subsystem includes a LED control panel, a camera and a computer, wherein
The computer is separately connected the LED control panel and the camera, and the LED control panel control connects the LED.
The present invention also provides a kind of eye movement real-time tracing algorithms, comprising:
(a) eye image of acquisition is adjusted to the resolution ratio of 320*256;
(b) image is inputted into network access network, by multilayer expansion convolution and the parallel operation of common convolution, extracts image respectively
The abstract characteristics of different levels;
(c) up-sampled by deconvolution and stack fusion, thus make each layer fusion after as a result, both comprising height
The abstract characteristics of image of degree includes the characteristics of image of local fine again;
(d) fused characteristics of image is passed through into active coating, carries out cost calculating with the Pupil Segmentation figure of original image, passes through
The loss that network feedforward is propagated is calculated, and gradient is calculated come adjusting parameter value by backpropagation;
(e) when network training iteration is to 60 times, according to the semantic segmentation figure of pupil, judge to calculate pupil mass center, according to this
Centroid position realizes the real-time location tracking of pupil, while the radius of pupil is calculated according to semantic segmentation figure.
The present invention also provides a kind of segmentations of macula lutea and center extraction algorithm, comprising:
(a) eye fundus image of acquisition is adjusted to the resolution ratio of 320*256, image is inputted into network access network, it is swollen by multilayer
Swollen convolution and the parallel operation of common convolution;
(b) up-sampled by deconvolution and stack fusion, thus make each layer fusion after as a result, both comprising height
The abstract characteristics of image of degree includes the characteristics of image of local fine again;
(c) fused characteristics of image is passed through into active coating, carries out cost calculating with the Pupil Segmentation figure of original image, passes through
The loss that network feedforward is propagated is calculated, and gradient is calculated come adjusting parameter value by backpropagation;
(d) when network training iteration is to 160 times, macular region more can be accurately partitioned into.
By adopting the above technical scheme, the beneficial effects of the present invention are:
(1) bright spot scintillation parameter be arranged, provide five kinds of bright spot colors, fourth gear bias light, bright spot flashing brilliant black ratio, seven kinds it is bright
The parameter setting functions such as point luminous intensity;
(2) critical radar cross section measurement and Drawing of Curve, record the test frequency data of tested personnel, detection terminates
Afterwards, data and curves can be provided, and are stored in database, are used for comparing;
(3) it monitors test process in real time, state is tracked by image capture module and eye movement, observe tested personnel and test shape
State;
(4) data base administration: storing and manage tested member's information, operator message, maintenance personnel's information, measurement parameter and
Analysis is as a result, the inquiry, backup and recovery of data can be carried out;
(5) data statistic analysis: carrying out individual to measurement parameter and community information statistically analyze, and generates fatigue conditions detection
Report, and support to print out;
(6) system maintaining module.
Detailed description of the invention
Fig. 1 is integrating sphere pictorial diagram of the invention and even smooth schematic diagram;
Fig. 2 is even light pipe pictorial diagram of the invention and even smooth schematic diagram;
Fig. 3 is abrasive segments pictorial diagram of the invention and even smooth schematic diagram;
Fig. 4 is electrical entire block diagram of the invention;
Fig. 5 is the schematic diagram at one of visual angle of amblyopia training instrument structure of the invention;
Fig. 6 is the schematic diagram at another visual angle of amblyopia training instrument structure of the invention;
Fig. 7 is customer service work flow diagram of the invention.
Specific embodiment
Specific embodiments of the present invention will be further explained with reference to the accompanying drawing.It should be noted that for
The explanation of these embodiments is used to help understand the present invention, but and does not constitute a limitation of the invention.In addition, disclosed below
The each embodiment of the present invention involved in technical characteristic can be combined with each other as long as they do not conflict with each other.
The present invention provides a kind of digitlization amblyopia training systems, comprising: a flicker fusion frequency measures subsystem, described
Flicker fusion frequency measuring system generates a visual imaging optical path;One computer control analysis subsystem, the computer control
It analyzes subsystem and controls current background light source, the computer control analysis subsystem control connects the flicker fusion frequency and surveys
Amount system;And one digitlization amblyopia training device.
Specifically, the digitlization amblyopia training device includes: an eyeshade;One visual channel, the visual channel setting
In the eyeshade, the visual imaging optical path is arranged in the visual channel;One pupil image component, pupil illumination at
As component is perpendicular to visual channel;One flashing bright spot component, the light-emitting surface of the flashing bright spot component is apart from the eyeshade
250mm, the computer control analysis subsystem control the flashing bright spot component;One reflecting mirror, the flashing bright spot component
Reflecting mirror described in the light-transmissive of sending enters pupil and stimulates macular region, and pupil is by the post-stimulatory variation feelings of bright spot
Condition will be in real time by the pupil image monitor component;One display screen, the result images of the pupil image component are shown in described
On display screen;And a host, the host control the flashing bright spot component and the flashing bright spot component and described show
Display screen.
The flicker fusion frequency measurement subsystem generates light source, and the computer control analysis subsystem receives and control
Light source, the digitlization amblyopia training device can control input, subject feedback's input and display output.
More specifically, the design considerations of the flicker fusion frequency measurement subsystem is as follows:
1) bright spot diameter: 2mm ± 0.2mm;
2) bright spot color: red, green, blue, Huang, white;
3) background luminous intensity: 1,1/4,1/16, it is completely black;
4) bright spot flicker frequency: 4~60Hz.
The flicker fusion frequency measuring system includes a source stimulating device and a monitoring camera, wherein the source stimulating
Device is provided with multiple bright spot light sources and background light source and a LED, and the source stimulating device uses the even light of integrating sphere, even light
One of means of the even light of light pipe or the even smooth means of abrasive segments realize the LED to the uniform illumination of the bright spot light source.
As shown in Figure 1, Figure 2 and Figure 3, according to design requirement, the even light of source stimulating device proposed adoption integrating sphere, even light pipe are even
Light or the even smooth means of abrasive segments realize the uniform illumination of LED to 2mm bright spot;Bias light is realized using circulating type LED illumination mode
It is effective uniformly;The design of visual imaging optical path proposed adoption total closed type visual channel, cooperates attaching type eyeshade, mentions for amblyopia training
For being completely independent controllable light environment.
More specifically, the design considerations of the computer control analysis subsystem is as follows:
1) electrical security: reference standard GB 9706.15-2008 " medical electrical equipment first part: want by safety general
Ask " regulation, guarantee the reliability and security of electrical system;
2) bright spot color: red, green, blue, Huang, white;
3) background luminous intensity: 1,1/4,1/16, it is completely black;
4) bright spot diameter: 2mm ± 0.2mm;
5) bright spot flicker frequency: 4~60Hz.
The computer control analysis subsystem includes a LED control panel, a camera and a computer, wherein the calculating
Machine is separately connected the LED control panel and the camera, and the LED control panel control connects the LED.
As shown in figure 4, PC computer is connected with LED control panel, camera respectively by the interface of two USB;LED control
Plate can adjust the flicker frequency of input control color LED, 5 kinds of colours of color adaptation input control according to external input, frequency
LED, duty ratio when bright spot duty cycle adjustment color LED flashing, the light intensity brightness of bright spot intensity adjustment color LED, backlight tune
Save the brightness of background White LED.
The structure design considerations of the digitlization amblyopia training device is as follows:
1) mechanical dimensions :≤400mm × 250mm × 350mm;
2) host weight :≤10kg;
3) environmental adaptability requirement:
A) operating temperature: 0 DEG C~40 DEG C;
B) storage temperature: -30 DEG C~55 DEG C;
C) working relative humidity: 10%-90%RH (non-condensing);
D) humidity: 5%-90%RH (non-condensing) is stored.
E) it vibrates: referring in GB/T 14710-2009 " medical electrical equipment environmental requirement and test method " 3.3 rule
It is fixed.
As shown in Figure 5 and Figure 6, digitlization amblyopia training instrument is by eyeshade, visual channel, host, display screen, pupil image group
The compositions such as part, flashing bright spot component, reflecting mirror, adjusting knob.Its flexibility eyeshade is by the research of ergonomics repeatedly and tests,
Human body face stickiness is good, reduces light leakage.Apart from human eye 250mm, pupil illuminates image-forming assembly and hangs down bright spot flickering unit light-emitting surface
Directly in visual channel.Bright spot issue light-transmissive ir reflector enter pupil and stimulate macular region, pupil by
The post-stimulatory situation of change of bright spot will be in real time by pupil image monitor component, and the results are shown on display screen.Rotation is adjusted
The flicker frequency and color of the adjustable bright spot of knob.
Program display screen provides touch operation function, and supports external connection keyboard input function, easy to repair and use, and one
Body modularized design conveniently carries, and it is good to repair characteristic.
Further, on the basis of the said structure, the present invention provides related algorithms, including following eye movement real-time tracing to calculate
Method and macula lutea segmentation and center extraction algorithm.
Specifically, a kind of eye movement real-time tracing algorithm is provided, comprising:
(a) eye image of acquisition is adjusted to the resolution ratio of 320*256;
(b) image is inputted into network access network, by multilayer expansion convolution and the parallel operation of common convolution, extracts image respectively
The abstract characteristics of different levels;
(c) up-sampled by deconvolution and stack fusion, thus make each layer fusion after as a result, both comprising height
The abstract characteristics of image of degree includes the characteristics of image of local fine again;
(d) fused characteristics of image is passed through into active coating, carries out cost calculating with the Pupil Segmentation figure of original image, passes through
The loss that network feedforward is propagated is calculated, and gradient is calculated come adjusting parameter value by backpropagation;
(e) when network training iteration is to 60 times, according to the semantic segmentation figure of pupil, judge to calculate pupil mass center, according to this
Centroid position realizes the real-time location tracking of pupil, while the radius of pupil is calculated according to semantic segmentation figure.
Specifically, a kind of segmentation of macula lutea and center extraction algorithm are provided, comprising:
(a) eye fundus image of acquisition is adjusted to the resolution ratio of 320*256, image is inputted into network access network, it is swollen by multilayer
Swollen convolution and the parallel operation of common convolution;
(b) up-sampled by deconvolution and stack fusion, thus make each layer fusion after as a result, both comprising height
The abstract characteristics of image of degree includes the characteristics of image of local fine again;
(c) fused characteristics of image is passed through into active coating, carries out cost calculating with the Pupil Segmentation figure of original image, passes through
The loss that network feedforward is propagated is calculated, and gradient is calculated come adjusting parameter value by backpropagation;
(d) when network training iteration is to 160 times, macular region more can be accurately partitioned into.
Since people's the structure of the eye is fixed, the center of macula lutea will not change for pupil center location,
Therefore according to above two segmentation [Pupil Segmentation and macula lutea are divided], it is established that the position mapping relations at two segmentation centers, system
Count the average coordinates deviation of 30 frames.The deviation of each frame calculates as follows:
Pupil coordinate is [x1, y1], and macula lutea coordinate is [x2, y2],
X-direction deviation [x2-x1]
Y-direction deviation [y2-y1]
Calculate the average value of 30 frames.So as to adjust bright spot position.
According to the systems and methods, functional effect of the invention is as follows:
(1) bright spot scintillation parameter be arranged, provide five kinds of bright spot colors, fourth gear bias light, bright spot flashing brilliant black ratio, seven kinds it is bright
The parameter setting functions such as point luminous intensity;
(2) critical radar cross section measurement and Drawing of Curve, record the test frequency data of tested personnel, detection terminates
Afterwards, data and curves can be provided, and are stored in database, are used for comparing;
(3) it monitors test process in real time, state is tracked by image capture module and eye movement, observe tested personnel and test shape
State;
(4) data base administration: storing and manage tested member's information, operator message, maintenance personnel's information, measurement parameter and
Analysis is as a result, the inquiry, backup and recovery of data can be carried out;
(5) data statistic analysis: carrying out individual to measurement parameter and community information statistically analyze, and generates fatigue conditions detection
Report, and support to print out;
(6) system maintaining module.
Specifically running environment and exploitation environment are as follows:
(1) running environment
A) CPU:Intel M3 series and the above performance;
B) memory: 2GB;
C) hard disk: 64G.
(2) environment is developed
A) operating system: Windows 7/10;
B) development system: Visual Studio 2013;
C) language: C++, C# is used.
As shown in fig. 7, administrator logs in software systems, a tested member or the tested member of registration are selected, then can be faced
The test of boundary's flicker fusion frequency, tested member watch bright spot attentively, when the bright spot flicker frequency of starting is lower, by adjusting frequency knob,
Gradually increase flicker frequency until feel bright spot no longer flash (or originate bright spot flicker frequency it is higher, gradually decrease sudden strain of a muscle from height
Bright frequency is until feel that bright spot flashes) it completes one and tests, then replaceable bright spot color, bright spot brightness, bias light
Continuing to test with parameters such as brilliant black ratios, test is completed to provide test result Drawing of Curve and saves data, in test process, software
The tested member's test mode of eye movement tracing monitoring is provided.Administrator can enter test pattern by login interface, tested without selecting
Member can carry out CFF test, and test pattern test data is saved without database.Administrator can also carry out repairing for bright spot parameter
Change, carries out different scheme tests.
In conjunction with attached drawing, the embodiments of the present invention are described in detail above, but the present invention is not limited to described implementations
Mode.For a person skilled in the art, in the case where not departing from the principle of the invention and spirit, to these embodiments
A variety of change, modification, replacement and modification are carried out, are still fallen in protection scope of the present invention.
Claims (8)
1. a kind of digitlization amblyopia training device, suitable for testing the pupil of training human eye characterized by comprising
One eyeshade;
One visual channel, the visual channel are arranged in the eyeshade;
One pupil image component, the pupil illumination image-forming assembly is perpendicular to visual channel;
One flashing bright spot component, the light-emitting surface of the flashing bright spot component is apart from the eyeshade 250mm;
One reflecting mirror, it is described flashing bright spot component issue light-transmissive described in reflecting mirror enter pupil and stimulate macular area
Domain, pupil is will be in real time by the pupil image monitor component by the post-stimulatory situation of change of bright spot;
The result images of one display screen, the pupil image component are shown on the display screen;And
One host, the host control the flashing bright spot component and the flashing bright spot component and the display screen.
2. digitlization amblyopia training device according to claim 1, which is characterized in that it further include an adjusting knob, it is described
Adjusting knob controls the flashing bright spot component, to adjust the flicker frequency and color of bright spot.
3. a kind of digitlization amblyopia training system characterized by comprising
One flicker fusion frequency measures subsystem, and the flicker fusion frequency measuring system generates a visual imaging optical path;
One computer control analysis subsystem, the computer control analysis subsystem control current background light source, the calculating
Machine control analysis subsystem control connects the flicker fusion frequency measuring system;And
One digitlization amblyopia training device, comprising:
One eyeshade;
One visual channel, the visual channel are arranged in the eyeshade, and the visual imaging optical path setting is logical in the vision
Road;
One pupil image component, the pupil illumination image-forming assembly is perpendicular to visual channel;
One flashing bright spot component, the light-emitting surface of the flashing bright spot component is apart from the eyeshade 250mm, the computer control point
It analyses subsystem and controls the flashing bright spot component;
One reflecting mirror, it is described flashing bright spot component issue light-transmissive described in reflecting mirror enter pupil and stimulate macular area
Domain, pupil is will be in real time by the pupil image monitor component by the post-stimulatory situation of change of bright spot;
The result images of one display screen, the pupil image component are shown on the display screen;And
One host, the host control the flashing bright spot component and the flashing bright spot component and the display screen.
4. digitlization amblyopia training system according to claim 3, which is characterized in that flicker fusion frequency measurement system
System includes a source stimulating device and a monitoring camera, wherein the source stimulating device is provided with multiple bright spot light sources and background light source
And a LED, and the source stimulating device using the even light of integrating sphere, the even light of even light pipe or the even smooth means of abrasive segments wherein
One means realizes the LED to the uniform illumination of the bright spot light source.
5. digitlization amblyopia training system according to claim 4, which is characterized in that the LED is circulating type LED.
6. digitlization amblyopia training system according to claim 4, which is characterized in that the computer control analysis, which divides, is
System includes a LED control panel, a camera and a computer, wherein the computer is separately connected the LED control panel and described
Camera, the LED control panel control connect the LED.
7. a kind of eye movement real-time tracing algorithm characterized by comprising
(a) eye image of acquisition is adjusted to the resolution ratio of 320*256;
(b) image is inputted into network access network, by multilayer expansion convolution and the parallel operation of common convolution, it is different extracts image respectively
The abstract characteristics of level;
(c) up-sampled by deconvolution and stack fusion, thus make each layer fusion after as a result, both comprising height take out
The characteristics of image of elephant includes the characteristics of image of local fine again;
(d) fused characteristics of image is passed through into active coating, carries out cost calculating with the Pupil Segmentation figure of original image, passes through calculating
The loss that network feedforward is propagated, and gradient is calculated come adjusting parameter value by backpropagation;
(e) when network training iteration is to 60 times, according to the semantic segmentation figure of pupil, judge to calculate pupil mass center, according to this mass center
The real-time location tracking of pupil is realized in position, while the radius of pupil is calculated according to semantic segmentation figure.
8. a kind of macula lutea segmentation and center extraction algorithm characterized by comprising
(a) eye fundus image of acquisition is adjusted to the resolution ratio of 320*256, image is inputted into network access network, expands and rolls up by multilayer
The long-pending and parallel operation of common convolution;
(b) up-sampled by deconvolution and stack fusion, thus make each layer fusion after as a result, both comprising height take out
The characteristics of image of elephant includes the characteristics of image of local fine again;
(c) fused characteristics of image is passed through into active coating, carries out cost calculating with the Pupil Segmentation figure of original image, passes through calculating
The loss that network feedforward is propagated, and gradient is calculated come adjusting parameter value by backpropagation;
(d) when network training iteration is to 160 times, macular region more can be accurately partitioned into.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910667755.XA CN110464611A (en) | 2019-07-23 | 2019-07-23 | A kind of digitlization amblyopia enhancing training device and system and its related algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910667755.XA CN110464611A (en) | 2019-07-23 | 2019-07-23 | A kind of digitlization amblyopia enhancing training device and system and its related algorithm |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110464611A true CN110464611A (en) | 2019-11-19 |
Family
ID=68508986
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910667755.XA Pending CN110464611A (en) | 2019-07-23 | 2019-07-23 | A kind of digitlization amblyopia enhancing training device and system and its related algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110464611A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112022641A (en) * | 2020-09-10 | 2020-12-04 | 深圳职业技术学院 | Method and system for assisting eye rotation movement |
CN112731668A (en) * | 2021-01-15 | 2021-04-30 | 东莞市东全智能科技有限公司 | Head-wearing type myopia adjusting device and myopia adjusting method |
Citations (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5861936A (en) * | 1996-07-26 | 1999-01-19 | Gillan Holdings Limited | Regulating focus in accordance with relationship of features of a person's eyes |
CN1871994A (en) * | 2006-06-29 | 2006-12-06 | 昆明钏泽智能***有限公司 | Digital system for detecting dynamic change of bilateral eye pupils |
CN204971326U (en) * | 2015-07-24 | 2016-01-20 | 中国人民解放军空军航空医学研究所 | Adopt critical flicker fusion frequency to detect tired device of maincenter |
CN106295139A (en) * | 2016-07-29 | 2017-01-04 | 姹ゅ钩 | A kind of tongue body autodiagnosis health cloud service system based on degree of depth convolutional neural networks |
CN107944399A (en) * | 2017-11-28 | 2018-04-20 | 广州大学 | A kind of pedestrian's recognition methods again based on convolutional neural networks target's center model |
CN108010049A (en) * | 2017-11-09 | 2018-05-08 | 华南理工大学 | Split the method in human hand region in stop-motion animation using full convolutional neural networks |
CN108416772A (en) * | 2018-03-07 | 2018-08-17 | 汕头大学 | A kind of strabismus detection method based on concatenated convolutional neural network |
CN108498064A (en) * | 2018-03-19 | 2018-09-07 | 中国人民解放军空军航空医学研究所 | Digitize bright spot scintillation detection system |
CN108717569A (en) * | 2018-05-16 | 2018-10-30 | 中国人民解放军陆军工程大学 | It is a kind of to expand full convolutional neural networks and its construction method |
CN109155053A (en) * | 2016-05-27 | 2019-01-04 | 索尼公司 | Information processing equipment, information processing method and recording medium |
CN109190752A (en) * | 2018-07-27 | 2019-01-11 | 国家新闻出版广电总局广播科学研究院 | The image, semantic dividing method of global characteristics and local feature based on deep learning |
CN109918752A (en) * | 2019-02-26 | 2019-06-21 | 华南理工大学 | Mechanical failure diagnostic method, equipment and medium based on migration convolutional neural networks |
CN109977812A (en) * | 2019-03-12 | 2019-07-05 | 南京邮电大学 | A kind of Vehicular video object detection method based on deep learning |
CN110009095A (en) * | 2019-03-04 | 2019-07-12 | 东南大学 | Road driving area efficient dividing method based on depth characteristic compression convolutional network |
CN111091524A (en) * | 2018-10-08 | 2020-05-01 | 天津工业大学 | Prostate transrectal ultrasound image segmentation method based on deep convolutional neural network |
-
2019
- 2019-07-23 CN CN201910667755.XA patent/CN110464611A/en active Pending
Patent Citations (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5861936A (en) * | 1996-07-26 | 1999-01-19 | Gillan Holdings Limited | Regulating focus in accordance with relationship of features of a person's eyes |
CN1871994A (en) * | 2006-06-29 | 2006-12-06 | 昆明钏泽智能***有限公司 | Digital system for detecting dynamic change of bilateral eye pupils |
CN204971326U (en) * | 2015-07-24 | 2016-01-20 | 中国人民解放军空军航空医学研究所 | Adopt critical flicker fusion frequency to detect tired device of maincenter |
CN109155053A (en) * | 2016-05-27 | 2019-01-04 | 索尼公司 | Information processing equipment, information processing method and recording medium |
CN106295139A (en) * | 2016-07-29 | 2017-01-04 | 姹ゅ钩 | A kind of tongue body autodiagnosis health cloud service system based on degree of depth convolutional neural networks |
CN108010049A (en) * | 2017-11-09 | 2018-05-08 | 华南理工大学 | Split the method in human hand region in stop-motion animation using full convolutional neural networks |
CN107944399A (en) * | 2017-11-28 | 2018-04-20 | 广州大学 | A kind of pedestrian's recognition methods again based on convolutional neural networks target's center model |
CN108416772A (en) * | 2018-03-07 | 2018-08-17 | 汕头大学 | A kind of strabismus detection method based on concatenated convolutional neural network |
CN108498064A (en) * | 2018-03-19 | 2018-09-07 | 中国人民解放军空军航空医学研究所 | Digitize bright spot scintillation detection system |
CN108717569A (en) * | 2018-05-16 | 2018-10-30 | 中国人民解放军陆军工程大学 | It is a kind of to expand full convolutional neural networks and its construction method |
CN109190752A (en) * | 2018-07-27 | 2019-01-11 | 国家新闻出版广电总局广播科学研究院 | The image, semantic dividing method of global characteristics and local feature based on deep learning |
CN111091524A (en) * | 2018-10-08 | 2020-05-01 | 天津工业大学 | Prostate transrectal ultrasound image segmentation method based on deep convolutional neural network |
CN109918752A (en) * | 2019-02-26 | 2019-06-21 | 华南理工大学 | Mechanical failure diagnostic method, equipment and medium based on migration convolutional neural networks |
CN110009095A (en) * | 2019-03-04 | 2019-07-12 | 东南大学 | Road driving area efficient dividing method based on depth characteristic compression convolutional network |
CN109977812A (en) * | 2019-03-12 | 2019-07-05 | 南京邮电大学 | A kind of Vehicular video object detection method based on deep learning |
Non-Patent Citations (2)
Title |
---|
景晨凯等: "基于深度卷积神经网络的人脸识别综述", 《计算机应用与软件》 * |
潘新星等: "基于卷积目标检测的3D眼球追踪***深度估计", 《仪器仪表学报》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112022641A (en) * | 2020-09-10 | 2020-12-04 | 深圳职业技术学院 | Method and system for assisting eye rotation movement |
CN112731668A (en) * | 2021-01-15 | 2021-04-30 | 东莞市东全智能科技有限公司 | Head-wearing type myopia adjusting device and myopia adjusting method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10386645B2 (en) | Digital therapeutic corrective spectacles | |
US10666918B2 (en) | Vision-based alerting based on physical contact prediction | |
US11102462B2 (en) | Vision defect determination via a dynamic eye characteristic-based fixation point | |
US7309128B2 (en) | Automated stereocampimeter and related method for improved measurement of the visual field | |
CN110573061B (en) | Ophthalmologic examination method and apparatus | |
US20190223716A1 (en) | Visual enhancement for dynamic vision defects | |
US20040156554A1 (en) | System and method for simulating visual defects | |
EP2236074A1 (en) | Visual display with illuminators for gaze tracking | |
US10742944B1 (en) | Vision defect determination for facilitating modifications for vision defects related to double vision or dynamic aberrations | |
CN201929941U (en) | Hemispheric stimulating vision function diagnosis and treatment instrument | |
Krösl et al. | Icthroughvr: Illuminating cataracts through virtual reality | |
US11039742B1 (en) | Vision testing via prediction-based setting of an initial stimuli characteristic for a user interface location | |
CN110464611A (en) | A kind of digitlization amblyopia enhancing training device and system and its related algorithm | |
JP2022525304A (en) | Visual defect determination and enhancement | |
Ma et al. | A one-step, streamlined children’s vision screening solution based on smartphone imaging for resource-limited areas: Design and preliminary field evaluation | |
CN102008286A (en) | Hemispherical projection-based vision function diagnosing and treating system | |
EP1872712B1 (en) | Device and method for a vision test | |
Schieber | Vision assessment technology and screening older drivers: Past practices and emerging techniques | |
CN110251073A (en) | It is a kind of intelligence screening strabismus and diopter diagnostic device | |
CN109875495A (en) | It is a kind of by display equipment the contactless full-automatic feedback visual field based on | |
CN214965416U (en) | Artificial artificial eye for replacing human eye for ophthalmologic equipment examination | |
RU2576798C2 (en) | Method for determining sight characteristics and device therefor | |
CN113080844A (en) | Visual detection and visual training device for optimizing retina area | |
Schubert | Encoding of the retinal image as observer-relative and object-relative luminance changes |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20191119 |