CN113075159B - Terahertz spectrum-based cell imaging system for slice inspection - Google Patents
Terahertz spectrum-based cell imaging system for slice inspection Download PDFInfo
- Publication number
- CN113075159B CN113075159B CN202110331741.8A CN202110331741A CN113075159B CN 113075159 B CN113075159 B CN 113075159B CN 202110331741 A CN202110331741 A CN 202110331741A CN 113075159 B CN113075159 B CN 113075159B
- Authority
- CN
- China
- Prior art keywords
- terahertz
- frequency domain
- sample
- terahertz frequency
- image processing
- 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.)
- Active
Links
- 238000001228 spectrum Methods 0.000 title claims abstract description 56
- 238000003384 imaging method Methods 0.000 title claims abstract description 39
- 238000007689 inspection Methods 0.000 title claims abstract description 16
- 238000012545 processing Methods 0.000 claims abstract description 45
- 238000013528 artificial neural network Methods 0.000 claims abstract description 18
- 238000000034 method Methods 0.000 claims abstract description 11
- 210000004881 tumor cell Anatomy 0.000 claims description 35
- 210000004027 cell Anatomy 0.000 claims description 26
- 238000007789 sealing Methods 0.000 claims description 26
- 238000001514 detection method Methods 0.000 claims description 24
- 238000012549 training Methods 0.000 claims description 24
- 239000007789 gas Substances 0.000 claims description 22
- 238000012795 verification Methods 0.000 claims description 10
- 238000010521 absorption reaction Methods 0.000 claims description 9
- CURLTUGMZLYLDI-UHFFFAOYSA-N Carbon dioxide Chemical compound O=C=O CURLTUGMZLYLDI-UHFFFAOYSA-N 0.000 claims description 8
- 239000004065 semiconductor Substances 0.000 claims description 8
- 238000013527 convolutional neural network Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 6
- 238000004611 spectroscopical analysis Methods 0.000 claims description 5
- 229910002092 carbon dioxide Inorganic materials 0.000 claims description 4
- 239000001569 carbon dioxide Substances 0.000 claims description 4
- JCXJVPUVTGWSNB-UHFFFAOYSA-N nitrogen dioxide Inorganic materials O=[N]=O JCXJVPUVTGWSNB-UHFFFAOYSA-N 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 3
- QJGQUHMNIGDVPM-UHFFFAOYSA-N nitrogen group Chemical group [N] QJGQUHMNIGDVPM-UHFFFAOYSA-N 0.000 claims description 2
- 206010028980 Neoplasm Diseases 0.000 abstract description 13
- 238000003745 diagnosis Methods 0.000 abstract description 10
- 238000012552 review Methods 0.000 abstract description 5
- 230000007306 turnover Effects 0.000 abstract description 3
- 239000000523 sample Substances 0.000 description 55
- 210000001519 tissue Anatomy 0.000 description 24
- 206010006187 Breast cancer Diseases 0.000 description 6
- 208000026310 Breast neoplasm Diseases 0.000 description 6
- 238000000862 absorption spectrum Methods 0.000 description 5
- 201000011510 cancer Diseases 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 210000004882 non-tumor cell Anatomy 0.000 description 4
- 230000001575 pathological effect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 238000012805 post-processing Methods 0.000 description 3
- 230000003595 spectral effect Effects 0.000 description 3
- 238000001328 terahertz time-domain spectroscopy Methods 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- IJGRMHOSHXDMSA-UHFFFAOYSA-N Atomic nitrogen Chemical compound N#N IJGRMHOSHXDMSA-UHFFFAOYSA-N 0.000 description 2
- 208000015634 Rectal Neoplasms Diseases 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 239000003153 chemical reaction reagent Substances 0.000 description 2
- 230000008602 contraction Effects 0.000 description 2
- 230000005855 radiation Effects 0.000 description 2
- 206010038038 rectal cancer Diseases 0.000 description 2
- 201000001275 rectum cancer Diseases 0.000 description 2
- NIXOWILDQLNWCW-UHFFFAOYSA-N acrylic acid group Chemical group C(C=C)(=O)O NIXOWILDQLNWCW-UHFFFAOYSA-N 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 239000012472 biological sample Substances 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 210000000481 breast Anatomy 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 210000000981 epithelium Anatomy 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000007490 hematoxylin and eosin (H&E) staining Methods 0.000 description 1
- 238000011532 immunohistochemical staining Methods 0.000 description 1
- 238000000338 in vitro Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 239000012188 paraffin wax Substances 0.000 description 1
- 231100000915 pathological change Toxicity 0.000 description 1
- 230000036285 pathological change Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/3581—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using far infrared light; using Terahertz radiation
Landscapes
- Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Health & Medical Sciences (AREA)
- Analytical Chemistry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Toxicology (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention provides a terahertz spectrum-based cell imaging system for slice inspection, which comprises a terahertz frequency domain spectrometer, a robot and control and image processing equipment; the terahertz frequency domain spectrometer is used for generating terahertz waves and collecting terahertz frequency domain spectrums generated after the sample is reflected by the terahertz waves; the robot is arranged beside the terahertz frequency domain spectrometer and used for clamping and overturning a sample; the control and image processing equipment is respectively connected with the terahertz frequency domain spectrometer and the robot and comprises a control subsystem and an image processing subsystem; the control subsystem is used for controlling the robot to clamp and turn over the sample and controlling the image processing subsystem to receive the terahertz frequency domain spectrum; the image processing subsystem uses an artificial neural network to perform three-dimensional imaging according to the terahertz frequency domain spectrum. The method can solve the technical problem that a doctor cannot review the diagnosis conclusion according to the terahertz time-domain spectrogram due to the fact that the tumor diagnosis conclusion given by the terahertz time-domain spectrogram can be misjudged due to external interference.
Description
Technical Field
The invention relates to the technical field of pathological examination, in particular to a section examination cell imaging system based on terahertz spectrum.
Background
The section examination is a kind of pathological examination, and is a pathomorphological method for examining pathological changes in body organs, tissues or cells, and is mainly used for pathological judgment of cancer at present. Specifically, HE staining and immunohistochemical staining are carried out after paraffin embedding, observation is carried out manually under an electron microscope, pathological judgment is made, the time is long, and the possibility of human misjudgment is high.
In the prior art, CN109540833A provides an in-vitro tumor tissue identification method based on a terahertz time-domain spectrum and application thereof, a specific absorption peak of the terahertz time-domain spectrum of a tumor tissue is utilized, and whether a tissue to be detected is a non-tumor tissue, a breast cancer tissue or a rectal cancer tissue is identified through comparison and analysis, biochemical reagents are not consumed in the whole detection process, the application of expensive reagents is avoided, and a large amount of manual operation is not needed. In order to realize the identification method, the tumor diagnosis system based on the terahertz time-domain spectroscopy comprises a detection module, a data processing module and a result output module, wherein the detection module is used for carrying out terahertz detection on a tissue sample to be detected; the data processing module is used for receiving the frequency domain information obtained by the detection module and calculating by using the frequency domain information to obtain a terahertz absorption spectrum; the result output module is used for receiving the terahertz absorption spectrum obtained from the data processing module, comparing the obtained terahertz absorption spectrum with the terahertz absorption spectrum of the tumor tissue, and judging whether the tissue sample to be detected is the breast cancer tissue, the rectal cancer tissue or the non-tumor tissue according to the comparison result.
According to the technical scheme, the terahertz absorption time-domain spectrum obtained by detection is compared with the terahertz absorption time-domain spectrum of the tumor tissue, and then the type of the tissue sample to be detected is directly judged according to the comparison result, so that the technical scheme has the following problems: when the detection module is used for carrying out terahertz detection on a tissue sample to be detected, the terahertz time-domain spectrum is interfered by external disturbance factors including air, power clutter and the like, and an obtained diagnosis conclusion may have errors. Such as: when the terahertz time-domain spectrum generated by the tissue sample to be detected is near the critical point of the terahertz absorption spectrum of the tumor tissue, for example, an absorption peak exists at 1.018THz, but due to interference of air and power supply noise, the frequency-domain information actually received by the data processing module is changed into 1.0185THz which is close to 1.019THz, at the moment, according to the technical scheme, the tumor diagnosis system of the terahertz time-domain spectrum can directly draw a conclusion, and the tissue sample to be detected is breast cancer tissue. Meanwhile, the tumor diagnosis system of the terahertz time-domain spectroscopy directly gives a diagnosis conclusion, and even if misjudgment occurs, a doctor cannot visually observe the tissue sample to be detected to make judgment again according to the terahertz time-domain spectroscopy, so that follow-up review of the diagnosis conclusion by the doctor is not facilitated.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a cell imaging system for section examination of a terahertz spectrum, which aims to solve the technical problem that a doctor cannot review a diagnosis conclusion according to a terahertz time-domain spectrogram due to the fact that the diagnosis conclusion of a tumor given by the terahertz time-domain spectrum is misjudged possibly because of the interference of external disturbance factors.
The technical scheme adopted by the invention is that a terahertz spectrum-based cell imaging system for slice inspection comprises a terahertz frequency domain spectrometer, a robot and control and image processing equipment;
the terahertz frequency domain spectrometer is used for generating terahertz waves and collecting terahertz frequency domain spectrums generated after the sample is reflected by the terahertz waves;
the robot is arranged beside the terahertz frequency domain spectrometer and used for clamping and overturning a sample;
the control and image processing equipment is respectively connected with the terahertz frequency domain spectrometer and the robot, and comprises a control subsystem and an image processing subsystem; the control subsystem is used for controlling the robot to clamp and turn over the sample and controlling the image processing subsystem to receive the terahertz frequency domain spectrum; the image processing subsystem uses an artificial neural network to perform three-dimensional imaging according to the terahertz frequency domain spectrum. The beneficial technical effects of the technical scheme are as follows: the three-dimensional images of high-resolution and high-contrast non-tumor cells, benign tumor cells and malignant tumor cells can be obtained, and a doctor can perform review and judgment according to the three-dimensional images.
In one implementation mode, the terahertz frequency domain spectrometer comprises two semiconductor lasers, a beam splitter, a bias device, a first mixer, a second mixer and a locking detection device;
the semiconductor laser is used for generating laser beams, the beam splitter is used for splitting the laser beams, the biasing device is used for applying biasing voltage to the first mixer, the first mixer and the second mixer are used for mixing, and the locking detection device is used for detecting frequency domain spectrums.
In one implementation, the control subsystem controls the robot to turn the sample 90 degrees in 1 degree steps as it turns the sample.
In one implementation, the image processing subsystem performs three-dimensional imaging according to the terahertz frequency domain spectrum using an artificial neural network, specifically as follows:
receiving a plurality of terahertz frequency domain spectrums, and then generating a plurality of two-dimensional images by using a gray scale modulation technology according to the difference value of the absorption coefficient and the refractive index in each terahertz frequency domain spectrum;
and generating a three-dimensional image by using an artificial neural network according to the plurality of two-dimensional images.
In one implementation, the training process of the artificial neural network is specifically as follows:
constructing a training set and a verification set according to the two-dimensional image;
training and verifying by using fast CNN according to the training set and the verification set to obtain a tumor cell detection network, wherein the output of the tumor cell detection network is a two-dimensional image of the tumor cell;
and taking the two-dimensional image of the tumor cell as input, and training and verifying by using C3D to obtain a tumor cell classification network.
In one implementation, the main parameters of Faster CNN are as follows:
the convolution kernel is 3 multiplied by 3, the number of the convolution kernels is 1024, the convolution layers are 3 layers, the activation function is ReLU, the full-connection layer is 1 layer, the learning rate is 0.01, and the iteration frequency is 10 ten thousand.
In one implementation, during training and verification by using C3D, features are extracted from a two-dimensional image of tumor cells by using 3D DPN using U-Net contraction/expansion, and then the features output by the 3D DPN and the original pixels of the two-dimensional image of tumor cells are provided as input to the GBM for classification.
In one implementation, the system operates as follows:
calibrating a terahertz frequency domain spectrometer;
clamping a sample by using a robot and placing the sample in a sample placing area of a terahertz frequency domain spectrometer;
controlling a terahertz frequency domain spectrometer to emit terahertz waves to a sample;
the control robot turns the sample by 90 degrees according to 1-degree stepping, and transmits the terahertz frequency domain spectrum received by the terahertz frequency domain spectrograph at each angle to an image processing subsystem of the control and image processing equipment;
and the image processing subsystem uses an artificial neural network to carry out three-dimensional imaging according to the terahertz frequency domain spectrum of the sample at each angle.
In an implementation mode, the system further comprises a transparent sealing cover, a vacuum pump and a gas storage tank, wherein the transparent sealing cover is respectively connected with the vacuum pump and the gas storage tank in a sealing manner; the terahertz frequency domain spectrometer and the robot are arranged in the sealing cover, and sealing parts are arranged at the positions, where the connecting wires of the terahertz frequency domain spectrometer, the robot and the control and image processing equipment penetrate through the sealing cover; the gas storage tank is filled with gas. The beneficial technical effects of the technical scheme are as follows: when a system generates a terahertz frequency domain spectrum by reflection of a sample, interference of water molecules in the air can be reduced, the accuracy of spectral data is greatly improved, and the three-dimensional imaging effect of cells is improved; and the adopted device has simple structure, low cost and convenient operation.
In one implementation, the gas contained in the gas container is nitrogen or carbon dioxide.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. Throughout the drawings, like elements or portions are generally identified by like reference numerals. In the drawings, elements or portions are not necessarily drawn to scale.
FIG. 1 is a schematic layout of a section examination cell imaging system based on terahertz spectroscopy according to embodiment 1 of the present invention;
FIG. 2 is a block diagram of a section inspection cell imaging system based on terahertz spectroscopy according to embodiment 1 of the present invention;
FIG. 3 is a block diagram of a terahertz frequency domain spectrometer system according to embodiment 1 of the present invention;
FIG. 4 is a graph showing the effect of cell imaging in example 1 of the present invention.
Reference numerals are as follows:
1-terahertz frequency domain spectrograph, 2-robot, 3-control and image processing equipment.
Detailed Description
Embodiments of the present invention will be described in detail below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and therefore are only used as examples, and the protection scope of the present invention is not limited thereby.
It is to be noted that, unless otherwise specified, technical or scientific terms used herein shall have the ordinary meaning as understood by those skilled in the art to which the invention pertains.
Example 1
The imaging system for the section inspection cell based on the terahertz spectrum provided by the embodiment comprises a terahertz frequency domain spectrometer 1, a robot 2 and a control and image processing device 3. The arrangement of the whole system in operation is shown in fig. 1, the block diagram is shown in fig. 2, the sample in the figure is not limited, and the sample is exemplified by a section made of a breast cancer tissue encapsulating slide, and the breast cancer tissue is a 200-micron breast epithelial tissue section with breast cancer cells.
The terahertz frequency domain spectrometer is shown in fig. 3 and comprises two semiconductor lasers, a beam splitter, a bias device, a first mixer, a second mixer and a locking detection device. Laser beams generated by the two semiconductor lasers are converged and then split by the beam splitter, one laser beam is radiated to the first mixer with the bias device to generate terahertz waves, the terahertz waves pass through the sample and reach the second mixer serving as the detector, and are converged with the other split laser beam, the two laser beams generate detectable current signals after frequency mixing, and then the terahertz frequency domain spectrum of the sample is obtained by detecting by using the locking detection device. In this embodiment, the semiconductor laser is a distributed feedback semiconductor laser, and the terahertz wave generated in the terahertz frequency domain spectrometer is narrowband continuous wave radiation.
In the embodiment, the terahertz frequency-domain spectrometer is used, and the terahertz waves generated by the frequency-domain spectrometer in use are continuous waves, which are different from pulse waves generated by the time-domain spectrometer, so that the continuous waves can obtain more comprehensive sample radiation information, and can obtain higher spectral resolution, which cannot be achieved by the time-domain spectrometer, and meanwhile, subsequent complex data processing is not needed, so that the terahertz frequency-domain spectrometer is very suitable for terahertz wave imaging of a sample.
In order to perform terahertz wave imaging on a sample, the terahertz wave is required to be used for carrying out multi-dimensional scanning on the sample, and post-processing imaging is carried out by using a control and image processing device according to the multi-dimensional scanning result. For carrying out the scanning of multidimension degree to the sample, use the robot to press from both sides the sample in this embodiment, overturn in the sample area of placing of terahertz frequency domain spectrum appearance now, the robot can select to use six industrial robot of desktop type. The robot operation, such as gripping, turning over, is realized by preprogramming, the programming mode is not limited, and the programming is realized by any realizable means in the prior art.
For different cell tissues, normal cells, benign tumor cells and malignant tumor cells in a sample, the absorption coefficient and the refractive index of the terahertz waves are different, so that when the sample is clamped by a robot and turned over in a sample placing area of the terahertz frequency domain spectrograph, the terahertz frequency domain spectrum of the sample obtained by detecting by a locking detection device is not completely consistent at each angle, and thus two-dimensional imaging can be performed according to the difference value of the absorption coefficient and the refractive index to obtain a two-dimensional image. In the embodiment, the terahertz frequency domain spectrometer detects a sample by using reflective measurement, so that when the angle of the sample is turned by 90 degrees, all two-dimensional images required by post-processing imaging can be obtained; if transmission measurement is adopted, the angle of the sample needs to be turned for 360 degrees to obtain all two-dimensional images required by post-processing imaging. In this embodiment, when the sample carries out the angle upset, through preprogramming to the robot, the upset angle that sets up the robot clamp and get the sample is step-by-step for 1 degree, adopts this kind of mode to obtain the terahertz frequency domain spectrum of 90 samples. The control and image processing equipment is connected with the terahertz frequency domain spectrograph, receives 90 terahertz frequency domain spectrums, and then generates 90 two-dimensional images by using a gray scale modulation technology according to the difference value of the absorption coefficient and the refractive index in each terahertz frequency domain spectrum.
In order to improve the resolution of the picture, the maximum contrast difference of all pixel points of the acquired two-dimensional image needs to be counted in advance, and gray level modulation is carried out after the maximum contrast is equally divided according to a certain numerical value; in this embodiment, it is preferable to perform 1000 equal divisions. And then, carrying out hierarchical modulation on the maximum contrast parameter by using a color grading method to obtain a two-dimensional image with color, high contrast and high resolution.
In order to facilitate the doctor to visually observe the sample, in this embodiment, the control and image processing device generates a three-dimensional image by using a two-dimensional image through an artificial neural network, and a training process of the artificial neural network is specifically as follows:
1. and collecting two-dimensional images, and constructing a training set and a verification set.
In this embodiment, before training the artificial neural network, a terahertz frequency domain spectrometer is used to collect multiple sets of two-dimensional images of multiple samples, where each set of each sample has 90 two-dimensional images with color, high contrast, and high resolution. Each group of images is used as a data sample, and a training set and a verification set are constructed by using the data samples; the data ratio of the training set and the validation set is 7.
2. And training and verifying by using fast CNN according to the training set and the verification set to obtain the tumor cell detection network.
In this embodiment, fast CNN (fast convolutional neural network) is selected for training. The Faster CNN uses the regional nomination network to directly calculate the candidate frame, so that the target detection speed can be obviously improved, and the positioning and classification accuracy of the target detection is higher. In this embodiment, to consider both the training efficiency and the classification accuracy of the network, the main parameters of the Faster CNN are as follows:
the convolution kernel is 3 multiplied by 3, the number of the convolution kernels is 1024, the convolution layers are 3 layers, the activation function is ReLU, the full-connection layer is 1 layer, the learning rate is 0.01, and the iteration frequency is 10 ten thousand. The fast CNN obtained through training is positioned as a tumor cell detection network in the embodiment, and the two-dimensional image can be a tumor cell two-dimensional image by inputting the two-dimensional image into the tumor cell detection network; the two-dimensional image of the tumor cell includes classification boundaries showing the tumor cell and the non-tumor cell, and position and size information in the sample.
3. And taking the two-dimensional image of the tumor cell as input, and training and verifying by using C3D to obtain a tumor cell classification network.
In this embodiment, C3D (convolutional neural network with three-dimensional convolutional kernel) is selected for training. C3D is a three-dimensional convolution kernel input by a multi-channel image, and the output is a three-dimensional characteristic, so that a three-dimensional image can be generated according to a two-dimensional image. Specifically, in this embodiment, a contraction/expansion structure of U-Net (semantic segmentation full convolution neural network) is adopted, a feature is extracted from a two-dimensional image of a tumor cell by using 3D DPN (dual path neural network with a convolution kernel being three-dimensional), and then the feature output by the 3D DPN and an original pixel of the two-dimensional image of the tumor cell are used as input and provided to GBM (gradient enhancement machine) for classification, so as to obtain a tumor cell classification network. The output result of the tumor cell classification network is a three-dimensional image, and the three-dimensional image comprises classification boundaries of non-tumor cells, benign tumor cells and malignant tumor cells, and position and size information in the sample.
In the step, the data sample of the verification set is obtained by constructing according to the judgment result after the two-dimensional image of the tumor cell is judged manually by a doctor.
The control and image processing device is an industrial personal computer in the embodiment, and is respectively and electrically connected with the robot and the terahertz frequency domain spectrometer, and a control subsystem and an image processing subsystem are installed on the industrial personal computer. The control subsystem can control the robot to clamp the sample and turn over the sample in the imaging process; the control subsystem can also control the work of the terahertz frequency domain spectrometer and control the image processing subsystem to receive the terahertz frequency domain spectrum. The image processing subsystem uses an artificial neural network to perform three-dimensional imaging according to the terahertz frequency domain spectrum.
The working principle of the embodiment 1 is explained in detail as follows:
1. calibrating terahertz frequency domain spectrograph
Firstly, the energy intensity and the frequency range of terahertz wave pulses are calibrated to ensure that the biological sample is irradiated under the condition of broadband and the same energy intensity of terahertz wave;
2. clamping a sample by using a robot and placing the sample in a sample placing area of a terahertz frequency domain spectrometer;
3. controlling a terahertz frequency domain spectrometer to emit terahertz waves to a sample;
4. the robot is controlled to turn the sample by 90 degrees according to 1-degree stepping, and the terahertz frequency domain spectrum received by the terahertz frequency domain spectrograph at each angle is transmitted to the image processing subsystem;
5. and the image processing subsystem uses an artificial neural network to carry out three-dimensional imaging according to the terahertz frequency domain spectrum of the sample at each angle. The three-dimensional image of the cell is shown in FIG. 4.
By the technical scheme of the embodiment, the three-dimensional images of the non-tumor cells, the benign tumor cells and the malignant tumor cells with high resolution and high contrast can be obtained, and a doctor can perform review and judgment according to the three-dimensional images.
Example 2
When a reflection terahertz frequency domain spectrometer is used for detecting a sample by using reflection type measurement, water molecules in air bring certain influence on the reflected wave when the reflected wave is reflected. In order to solve the technical problems, the following technical scheme is adopted for further optimization on the basis of the embodiment 1:
the terahertz spectrum based cell imaging system for slice inspection further comprises a transparent sealing cover, a vacuum pump and a gas storage tank, wherein the transparent sealing cover is respectively in sealing connection with the vacuum pump and the gas storage tank; the terahertz frequency domain spectrometer and the robot are arranged in the sealing cover, and sealing parts are arranged at the positions, where the connecting wires of the terahertz frequency domain spectrometer, the robot and the control and image processing equipment penetrate through the sealing cover; the gas storage tank is filled with gas.
The working principle of example 1 is explained in detail below:
in this embodiment, the transparent sealing cover is made of, but not limited to, acrylic, and has a length of 2 meters, a width of 1 meter, and a height of 1 meter. The terahertz frequency domain spectrometer and the robot are arranged in the sealing cover, and the control and image processing equipment is arranged outside the sealing cover. And a rubber ring is selected as a sealing element at the position where the connecting wires of the terahertz frequency domain spectrometer, the robot and the control and image processing equipment penetrate through the sealing cover.
An operator first performs a vacuum process on the internal gas of the transparent sealing cover by using a vacuum pump. The extraction time is determined according to the volume of the transparent sealing cover and the extraction amount of the vacuum pump, when the vacuum pump extracts vacuum according to the extraction time obtained through calculation, an operator opens the valve of the gas storage tank, and gas in the gas storage tank can rapidly enter the transparent sealing cover under the action of negative pressure. In this embodiment, the gas contained in the gas tank may be nitrogen or carbon dioxide. When the transparent sealing cover is filled with nitrogen or carbon dioxide, the section examination cell imaging system based on the terahertz spectrum is controlled to work according to the technical scheme of the embodiment 1.
By the technical scheme of the embodiment, when the system generates the terahertz frequency domain spectrum by the reflection of the sample, the interference of water molecules in the air can be reduced, the accuracy of spectral data is greatly improved, and the three-dimensional imaging effect of cells is improved; and the adopted device has simple structure, low cost and convenient operation.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; such modifications and substitutions do not depart from the spirit and scope of the present invention, and they should be construed as being included in the following claims and description.
Claims (9)
1. A section inspection cell imaging system based on terahertz spectrum which is characterized in that: the terahertz frequency domain spectrometer comprises a terahertz frequency domain spectrometer, a robot and control and image processing equipment;
the terahertz frequency domain spectrometer is used for generating terahertz waves and collecting terahertz frequency domain spectrums generated after a sample is reflected by the terahertz waves;
the robot is arranged beside the terahertz frequency domain spectrometer and used for clamping and overturning a sample;
the control and image processing equipment is respectively connected with the terahertz frequency domain spectrograph and the robot, and comprises a control subsystem and an image processing subsystem; the control subsystem is used for controlling the robot to clamp the sample and turning the sample by 90 degrees according to 1-degree stepping; the image processing subsystem is also used for controlling the image processing subsystem to receive the terahertz frequency domain spectrum of the sample under each angle; the image processing subsystem uses an artificial neural network to perform three-dimensional imaging according to the terahertz frequency domain spectrum.
2. The system for imaging slicing inspection cells based on terahertz spectrum as claimed in claim 1, wherein the terahertz frequency domain spectrometer comprises two semiconductor lasers, a beam splitter, a biasing device, a first mixer, a second mixer, a lock detection device;
the semiconductor laser is used for generating laser beams, the beam splitter is used for splitting laser beams, the bias device is used for applying bias voltage to the first mixer, the first mixer and the second mixer are used for mixing, and the locking detection device is used for detecting frequency domain spectrums.
3. The system for imaging a slice inspection cell based on terahertz spectrum according to claim 1, wherein the image processing subsystem performs three-dimensional imaging according to terahertz frequency domain spectrum by using an artificial neural network, and comprises the following components:
receiving a plurality of terahertz frequency domain spectrums, and then generating a plurality of two-dimensional images by using a gray scale modulation technology according to the difference value of the absorption coefficient and the refractive index in each terahertz frequency domain spectrum;
and generating a three-dimensional image by using an artificial neural network according to the plurality of two-dimensional images.
4. The system for imaging slice inspection cells based on terahertz spectroscopy as claimed in claim 3, wherein the training process of the artificial neural network is as follows:
constructing a training set and a verification set according to the two-dimensional image;
training and verifying by using fast CNN according to the training set and the verification set to obtain a tumor cell detection network, wherein the output of the tumor cell detection network is a two-dimensional image of the tumor cell;
and taking the two-dimensional image of the tumor cell as input, and training and verifying by using C3D to obtain a tumor cell classification network.
5. The terahertz spectroscopy-based slice inspection cell imaging system of claim 4, wherein the main parameters of the Faster CNN are as follows:
the convolution kernel is 3 multiplied by 3, the number of the convolution kernels is 1024, the convolution layers are 3 layers, the activation function is ReLU, the full-connection layer is 1 layer, the learning rate is 0.01, and the iteration frequency is 10 ten thousand.
6. The terahertz spectrum-based slice inspection cell imaging system according to claim 4, wherein: when the C3D is used for training and verification, the shrinkage/expansion of U-Net is adopted to extract features from the two-dimensional image of the tumor cell by using 3D DPN, and then the features output by the 3D DPN and the original pixels of the two-dimensional image of the tumor cell are used as input and are provided for GBM to be classified.
7. The terahertz spectroscopy-based slice inspection cell imaging system of claim 1, wherein the system operates by:
calibrating a terahertz frequency domain spectrometer;
clamping a sample by using a robot and placing the sample in a sample placing area of a terahertz frequency domain spectrometer;
controlling a terahertz frequency domain spectrometer to emit terahertz waves to a sample;
the control robot turns the sample by 90 degrees according to 1-degree stepping, and transmits the terahertz frequency domain spectrum received by the terahertz frequency domain spectrograph at each angle to an image processing subsystem of the control and image processing equipment;
and the image processing subsystem uses the artificial neural network to carry out three-dimensional imaging according to the terahertz frequency domain spectrum of the sample at each angle.
8. The system for imaging cells for slice inspection based on terahertz spectrum according to claim 1, further comprising a transparent sealing cover, a vacuum pump and a gas storage tank, wherein the transparent sealing cover is hermetically connected with the vacuum pump and the gas storage tank respectively; the terahertz frequency domain spectrometer and the robot are arranged in the sealing cover, and sealing parts are arranged at the positions, where the connecting wires of the terahertz frequency domain spectrometer, the robot and the control and image processing equipment penetrate through the sealing cover; the gas storage tank is filled with gas.
9. The terahertz spectrum-based slice inspection cell imaging system as set forth in claim 1, wherein: the gas contained in the gas storage tank is nitrogen or carbon dioxide.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110331741.8A CN113075159B (en) | 2021-03-26 | 2021-03-26 | Terahertz spectrum-based cell imaging system for slice inspection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110331741.8A CN113075159B (en) | 2021-03-26 | 2021-03-26 | Terahertz spectrum-based cell imaging system for slice inspection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113075159A CN113075159A (en) | 2021-07-06 |
| CN113075159B true CN113075159B (en) | 2022-10-14 |
Family
ID=76611252
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110331741.8A Active CN113075159B (en) | 2021-03-26 | 2021-03-26 | Terahertz spectrum-based cell imaging system for slice inspection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113075159B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114062302A (en) * | 2021-09-27 | 2022-02-18 | 国网河北省电力有限公司雄安新区供电公司 | Distribution network autonomous inspection method for terahertz imaging detection |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108267419A (en) * | 2017-12-08 | 2018-07-10 | 山东省科学院自动化研究所 | The method that terahertz time-domain spectroscopy detects adhesive bonding of composites structure debonding defect |
| CN110243827A (en) * | 2019-07-18 | 2019-09-17 | 华中科技大学 | A kind of rapid three dimensional imaging process suitable for light transparent sample |
| CN110585317A (en) * | 2019-10-25 | 2019-12-20 | 重庆医药高等专科学校 | Processing method and process screening method of rhizoma pinellinae praeparata |
| CN112446335A (en) * | 2020-12-02 | 2021-03-05 | 电子科技大学中山学院 | Terahertz contraband detection method based on deep learning |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AUPR872901A0 (en) * | 2001-11-09 | 2001-11-29 | Marine Research Wa Pty Ltd | Improved real or near real time earth imaging system |
| US20180120196A1 (en) * | 2016-10-31 | 2018-05-03 | The Boeing Company | Method and system for non-destructive testing using an unmanned aerial vehicle |
| EP3363357A1 (en) * | 2017-02-20 | 2018-08-22 | Koninklijke Philips N.V. | Contactless skin surface conductivity detector |
| US10234383B2 (en) * | 2017-06-20 | 2019-03-19 | Konica Minolta Laboratory U.S.A., Inc. | Terahertz spectral imaging system and security surveillance system employing the same |
| CN112240880B (en) * | 2020-09-18 | 2021-10-08 | 华中科技大学 | Super-resolution stimulated Raman microscopic imaging method and device for realizing near resonance enhancement |
-
2021
- 2021-03-26 CN CN202110331741.8A patent/CN113075159B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108267419A (en) * | 2017-12-08 | 2018-07-10 | 山东省科学院自动化研究所 | The method that terahertz time-domain spectroscopy detects adhesive bonding of composites structure debonding defect |
| CN110243827A (en) * | 2019-07-18 | 2019-09-17 | 华中科技大学 | A kind of rapid three dimensional imaging process suitable for light transparent sample |
| CN110585317A (en) * | 2019-10-25 | 2019-12-20 | 重庆医药高等专科学校 | Processing method and process screening method of rhizoma pinellinae praeparata |
| CN112446335A (en) * | 2020-12-02 | 2021-03-05 | 电子科技大学中山学院 | Terahertz contraband detection method based on deep learning |
Non-Patent Citations (2)
| Title |
|---|
| Measuring the topological charge of terahertz cortex beams with a focal hyperbolic lens;Yang YQ等;《APPLIED OPTICS》;20200520;第4685-4691页 * |
| 抗生素类药物的太赫兹光谱自动识别***研究;王婷等;《电子世界》;20170723;第175-189页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113075159A (en) | 2021-07-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10921266B2 (en) | Imaging device, imaging method, and imaging system | |
| US4191940A (en) | Method and apparatus for analyzing microscopic specimens and the like | |
| RU2372610C1 (en) | Method and device for radiation monitoring liquid objects | |
| CN107144546B (en) | Terahertz imaging method based on reflected time domain waveform windowing | |
| DE112012001926T5 (en) | Method and device for detecting a hidden dangerous substance | |
| AU2012272604A1 (en) | Method for analyzing biological specimens by spectral imaging | |
| CN113075159B (en) | Terahertz spectrum-based cell imaging system for slice inspection | |
| CN110308108A (en) | Content of baicalin detection method and system based on terahertz time-domain spectroscopic technology | |
| CN113155809A (en) | Novel spectral detection method for ore classification and real-time quantitative analysis | |
| CN111289496A (en) | Detection method and device for long-distance zoom laser-induced breakdown spectroscopy | |
| CN106483152B (en) | A kind of X-ray energy spectrum imaging method | |
| CN107976408A (en) | A kind of historical relic non-destructive testing device | |
| CN111272735B (en) | Detection method of laser-induced breakdown spectroscopy | |
| CN110751048B (en) | Laser probe classification method and device based on image characteristic automatic spectral line selection | |
| CN112712877A (en) | Large-view-field high-flux high-resolution pathological section analyzer | |
| CN106796144A (en) | Method for determining absorption band | |
| CN114689598A (en) | Terahertz wave-based internal defect imaging method, electronic device, and storage medium | |
| CN111624194B (en) | Blade element nondestructive measurement method based on laser-induced breakdown spectroscopy | |
| CN209485976U (en) | A kind of dedicated infrared transflector spectral measurement attachment of glass slide sample preparation | |
| CN208621514U (en) | A kind of THz wave computer-aided tomography imaging device | |
| CN107421639B (en) | Plasma three-dimensional information diagnosis system in material surface treatment process | |
| CN112683859B (en) | Component quantitative analysis method, test system and storage medium | |
| CN109490239A (en) | A kind of dedicated infrared transflector spectral measurement attachment of glass slide sample preparation | |
| Xu et al. | Materials Identification of Polarized Pulse Laser Detection Based on Sparse Autoencoder and Softmax Classifier Framework | |
| KR101856167B1 (en) | Method of inducing probability density function of wood species, nondestructive method of analyzing wood species and nondestructive device of analyzing wood species |
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 |