CN110693458B - Intraoperative real-time parathyroid gland identification method based on near-infrared autofluorescence - Google Patents
Intraoperative real-time parathyroid gland identification method based on near-infrared autofluorescence Download PDFInfo
- Publication number
- CN110693458B CN110693458B CN201911016165.7A CN201911016165A CN110693458B CN 110693458 B CN110693458 B CN 110693458B CN 201911016165 A CN201911016165 A CN 201911016165A CN 110693458 B CN110693458 B CN 110693458B
- Authority
- CN
- China
- Prior art keywords
- parathyroid gland
- time
- parathyroid
- gland
- suspicious
- 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
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/0059—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence
- A61B5/0071—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence by measuring fluorescence emission
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/0059—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence
- A61B5/0075—Measuring for diagnostic purposes; Identification of persons using light, e.g. diagnosis by transillumination, diascopy, fluorescence by spectroscopy, i.e. measuring spectra, e.g. Raman spectroscopy, infrared absorption spectroscopy
-
- 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
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Molecular Biology (AREA)
- Medical Informatics (AREA)
- Artificial Intelligence (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Evolutionary Computation (AREA)
- Signal Processing (AREA)
- Psychiatry (AREA)
- Physiology (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Fuzzy Systems (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
Abstract
The invention relates to the technical field of parathyroid gland identification, in particular to an intraoperative real-time parathyroid gland identification method based on near-infrared autofluorescence, which comprises the following steps: imaging a suspected parathyroid gland; obtaining fluorescence spectrum data of suspicious parathyroid gland and surrounding tissues; judging whether the spectrogram is parathyroid gland; carrying out big data analysis; in the near-infrared autofluorescence-based intraoperative real-time parathyroid gland identification method, clinical medicine, photoelectronic technology and artificial intelligence are combined through research of a near-infrared autofluorescence-based parathyroid gland intelligent identification platform, various subjective and objective interference factors are eliminated to the maximum extent, real-time, objective, noninvasive, contrast agent-free and intelligent convenient parathyroid gland positioning identification in an operation is achieved, the maximum problem of thyroid surgery is solved, an objective and rapid parathyroid gland positioning identification method is established, and the method has great significance for development of the whole thyroid surgery.
Description
Technical Field
The invention relates to the technical field of parathyroid gland identification, in particular to an intraoperative real-time parathyroid gland identification method based on near-infrared autofluorescence.
Background
In recent years, the incidence rate of thyroid tumors in the world rapidly rises, the incidence rate of thyroid cancers in some provinces and cities of China already enters the first ten malignant tumor incidence rates, and the number of operations is rapidly increased. Parathyroid glands must be protected during thyroid surgery to avoid HypoPT has become a surgical consensus. However, the parathyroid gland is difficult to accurately locate and identify in real time, which is the most common and difficult problem to solve in the current thyroid surgery. In view of the above, we propose an intraoperative real-time parathyroid gland identification method based on near-infrared autofluorescence.
Disclosure of Invention
The invention aims to solve the defect that the parathyroid gland proposed in the background art is difficult to accurately position and identify in real time by proposing a near infrared autofluorescence-based intraoperative real-time parathyroid gland identification method.
The technical scheme adopted by the invention is as follows: the method comprises the following steps:
s1, imaging the suspicious parathyroid gland by adopting parathyroid gland imaging equipment;
s2, obtaining fluorescence spectrum data of the suspicious parathyroid gland and surrounding tissues by adopting parathyroid gland spectrum distinguishing equipment;
s3, judging whether the spectrogram is parathyroid gland by adopting an intraoperative freezing principle;
s4, collecting spectrogram data of parathyroid gland, carrying out big data analysis, and forming an artificial intelligence model;
and S5, forming parathyroid gland distinguishing equipment by using a trainer intelligent model, and inputting the imaging picture of the suspicious parathyroid gland into the parathyroid gland distinguishing equipment to judge whether the parathyroid gland is the parathyroid gland.
As a preferred technical scheme of the invention: in the step S1, the parathyroid gland imaging device comprises a 785nm laser area light source, a 3CCD high-sensitivity imaging system and a filtering system, and the filtering system comprises a 808nm long-pass filter and a 822nm narrow-band filter.
As a preferred technical scheme of the invention: in S2, the parathyroid gland spectrum distinguishing device includes a spectrometer and a 785nm laser point light source.
As a preferred technical scheme of the invention: the parathyroid gland spectrum distinguishing device in the S2 comprises the following specific steps:
s2.1, obtaining fluorescence spectrograms of suspicious tissues, thyroid, fat and lymph nodes;
and S2.2, judging whether the parathyroid gland of the suspicious tissue is determined according to the difference of fluorescence peak intensities by taking the fluorescence intensities of the fat and the normal thyroid gland as a reference.
As a preferred technical scheme of the invention: in the S4, the big data analysis method selects Fisher algorithm, Svm algorithm or Knn algorithm.
As a preferred technical scheme of the invention: the Fisher algorithm adopts linear discriminant analysis, and considers the inter-class information of the training sample on the basis of using a PCA method to reduce the dimension.
As a preferred technical scheme of the invention: the Svm algorithm adopts a kernel function to carry out dimension increasing on data, adopts different functions as kernel functions K (x, xi), constructs a learning machine for realizing input of different types of nonlinear decision surfaces, and comprises the following postures:
posture one: a polynomial kernel function, which has the formula:
K(x,xi)＝[(x*xi)+1]q；
the result is a polynomial classifier of order q;
and (5) attitude II: the radial basis kernel function has the following formula:
posture three: sigmoid kernel, whose formula is as follows:
K(x,xi)＝tanh(v(x*xi)+c)。
as a preferred technical scheme of the invention: the Knn algorithm is defined as: defining a discriminant function as &i(x)＝kiI 1,2, c, where ki represents the number of samples belonging to ω i class in k nearest neighbors, and the decision rule is that if gi (x) maxki, x ∈ ω i.
Compared with the prior art, the invention has the beneficial effects that: the research of the parathyroid gland intelligent identification platform based on near infrared autofluorescence combines clinical medicine with photoelectronic technology and artificial intelligence, eliminates various subjective and objective interference factors to the maximum extent, realizes real-time, objective, noninvasive, contrast-free, intelligent and convenient positioning identification of parathyroid gland in operation, solves the biggest difficult problem of thyroid surgery, establishes a method for objective, rapid positioning identification of parathyroid gland, and has great significance for the development of the whole thyroid surgery.
Drawings
FIG. 1 is an overall flow diagram of a preferred embodiment of the present invention;
FIG. 2 is a diagram of a parathyroid gland imaging apparatus mode of operation of a preferred embodiment of the present invention;
FIG. 3 is a diagram of the operating mode of a parathyroid gland spectrum discrimination apparatus in accordance with a preferred embodiment of the present invention;
FIG. 4 is a Fisher linear discriminant diagram according to a preferred embodiment of the present invention.
Detailed Description
It should be noted that, in the present application, features of embodiments and embodiments may be combined with each other without conflict, and technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1 to 4, a preferred embodiment of the present invention provides a near-infrared autofluorescence-based intraoperative real-time parathyroid gland identification method, which comprises the following steps:
s1, imaging the suspicious parathyroid gland by adopting parathyroid gland imaging equipment;
s2, obtaining fluorescence spectrum data of the suspicious parathyroid gland and surrounding tissues by adopting parathyroid gland spectrum distinguishing equipment;
s3, judging whether the spectrogram is parathyroid gland by adopting an intraoperative freezing principle;
s4, collecting spectrogram data of parathyroid gland, carrying out big data analysis, and forming an artificial intelligence model;
and S5, training an intelligent model to form parathyroid gland distinguishing equipment, inputting the imaging picture of the suspicious parathyroid gland into the parathyroid gland distinguishing equipment, and distinguishing whether the imaging picture is the parathyroid gland.
In the embodiment, parathyroid gland imaging is based on the principle that laser-induced autofluorescence intensity of parathyroid gland is obviously higher than that of thyroid gland, lymph node and fat, a parathyroid gland imaging system with high sensitivity and high signal-to-noise ratio is to be constructed, early-stage research proves that the autofluorescence peak wavelength of 785nm laser-induced parathyroid gland and thyroid gland is near 822nm, the parathyroid gland imaging system is to utilize 785nm laser-induced operation area fluorescence, a CCD imaging system with high sensitivity and high signal-to-noise ratio is used, and a self-designed filtering system and image fusion software are matched to receive fluorescence signals, so that real-time parathyroid gland near-infrared imaging in an operation is realized.
Further, in S1, the parathyroid gland imaging device includes a 785nm laser surface light source, a 3CCD high-sensitivity imaging system and a filter system, and the parathyroid gland imaging device includes a 785nm laser surface light source; the filter system comprises a 808nm long-pass filter and a 822nm narrow-band filter which are processed and tested by a light machine and are used for filtering exciting light of 785nm and allowing autofluorescence to pass; the 3CCD high-sensitivity imaging system is preferably a KarlStorz3CCD high-sensitivity imaging system. In S2, parathyroid gland spectrum distinguishing devices including a spectrometer and a 785nm laser point light source are purchased commercially; the optical fiber probe including the output receiving port is adaptively modified by the optical machine, and the adjustable fixing support is constructed with the assistance of the optical machine.
The near-infrared laser excitation tissue autofluorescence capable of imaging needs a certain power density, the power density is related to the power of a laser and the size of a laser spot (the power density is output power/spot area), the spot is large, the range of one-time detection is enlarged, but the power density is too small to excite fluorescence to be weak, so that detection is affected; the light spot is small, the range which can be detected at one time is reduced, the operation times are increased, the operation time is increased, and the tissue can be burnt if the power density is too high. When the spectral analysis is carried out, the tissue can be burnt due to overlarge exciting light power; autofluorescence induced by too little power is too weak, and autofluorescence differences of different tissues can be difficult to show; meanwhile, the laser pulse time has corresponding influence on the generation of autofluorescence.
In addition, when the paranoid gland imaging equipment is used, firstly, before the test is started, an operating room light source must be turned off to eliminate light interference, then, an operating area is scanned to obtain a near infrared image of the approximate position of a suspicious parathyroid gland, then, the operating room light is turned on, a visible light image is shot at the same position, and the image is fused by using an image registration fusion algorithm to obtain the parathyroid gland position information.
Specifically, a special image registration fusion algorithm is written by an optical machine. The medical image is a multi-mode image, and the registration method mainly uses a feature-based registration method. The feature-based registration method can obtain the significant features of the image, reduce the information amount of the image and reduce the calculation amount and errors of registration. A multi-purpose SelfQuotientimage or DOG method in the field of face recognition is adopted, images are processed through a multi-mode filtering method to eliminate the influence of modes, the remarkable characteristics of the images are highlighted, the characteristic selection effect of the images is improved, the influence of a multi-mode light source is eliminated, and the images are kept consistent.
It is worth to say that the image registration fusion algorithm has the following steps:
(1) the acquired image is subjected to a data preprocessing stage to obtain an image which can be used for segmentation, the preprocessing stage is used for denoising the image through a median filtering algorithm, and the contrast of the image is enhanced through a histogram equalization method;
(2) filtering modal influence by a multi-modal filtering method;
(3) respectively finding characteristic points of the near infrared image and the visible light image by using SURF characteristics;
(4) matching feature points are found based on the segmentation and the images are registered.
According to previous research, the laser power density of 0.4-1W/cm2 can excite detectable autofluorescence. The power of a commercially available 785nm laser in China is 3W, and the selectable spot area is calculated to be 7.5-3cm2, namely the spot diameter is 2.7-1.7 cm. Considering that the number of operations is reduced as far as possible under the premise of achieving the imaging effect in the actual operation, the invention is to select two light spot diameters of 2.5cm and 2 cm.
The intensity of autofluorescence is relatively weak, the distance of the camera of the imaging system has a determining function on whether an effective signal can be obtained, and the data of the previous research shows that the camera is generally positioned at a position 15-20cm away from the operative field and is perpendicular to the operative field as much as possible to image.
In this embodiment, the specific steps of the parathyroid gland spectrum distinguishing device in S2 are as follows:
s2.1, obtaining fluorescence spectrograms of suspicious tissues, thyroid, fat and lymph nodes;
and S2.2, judging whether the parathyroid gland of the suspicious tissue is determined according to the difference of fluorescence peak intensities by taking the fluorescence intensities of the fat and the normal thyroid gland as a reference.
Specifically, the spectrum and data were obtained by irradiating light at 80mW for 1 second. The test sequence was as follows: fat and normal thyroid tissue (the two tissues can be confirmed by 100% naked eyes) are firstly tested as benchmark, then suspicious parathyroid gland tissue and definite lymph node are tested, and parathyroid gland is judged manually before the standard that the fluorescence peak intensity on a spectrogram exceeds 1 time. Then all tested tissues are cut into small pieces and sent to the frozen pathology, and the previous judgment is verified according to the pathological gold standard.
In this embodiment, the Fisher algorithm, Svm algorithm or Knn algorithm is selected as the big data analysis method.
Among them, Fisher linear discriminant is also called linear discriminant analysis. FLD is an efficient method for global feature extraction based on sample classes. The method considers the inter-class information of training samples on the basis of using a PCA method to reduce the dimension, and the FLD has the basic principle that an optimal projection axis is found, so that the distance between the projections of various samples on the axis is as far as possible, and the projection of the samples in each class is as compact as possible, thereby the classification effect is optimal, namely, the inter-class distance is maximized and the intra-class distance is minimized. The FLD method has wide application in the aspect of image overall feature extraction. When applying statistical methods to solve the pattern recognition problem, the so-called "dimensional disaster" problem is often encountered, and methods that are applicable in low-dimensional spaces may not be applicable at all in high-dimensional spaces. It is sometimes important to compress the dimensions of the feature space. The Fisher method actually involves the problem of dimension compression. It is mathematically easy to compress a feature space into one dimension if points of the multi-dimensional feature space are projected onto a straight line. However, samples that are easily separated in a high dimensional space, projected onto any one line, may be mixed and indistinguishable from different classes of samples, as shown in FIG. 4(a) projected onto the xl or x2 axes. If the straight line is rotated around the origin, it is possible to find a direction in which the samples are projected onto the straight line, and the various types of samples can be well separated, as shown in fig. 4 (b). The choice of the direction of the line is therefore important, and in general a best direction can always be found, making it easy to separate the line in which the sample is projected. How to find this best straight direction and how to implement the transformation of the projection to the best direction, the Fisher algorithm is used to solve this basic problem.
The Svm algorithm adopts a kernel function to carry out dimension increasing on data, adopts different functions as the kernel function K (x, xi), constructs a learning machine for realizing input of different types of nonlinear decision surfaces, and comprises the following postures:
posture one: a polynomial kernel function, whose formula is as follows:
K(x,xi)＝[(x*xi)+1]q；
the result is a polynomial classifier of order q;
and (5) posture II: the radial basis kernel function has the following formula:
posture three: sigmoid kernel, whose formula is as follows:
K(x,xi)＝tanh(v(x*xi)+c)。
the SVM has the following advantages:
(1) the method is specially used for the limited sample condition, and aims to obtain the optimal value under the existing sample information, not only the optimal value when the number of samples tends to infinity;
(2) the algorithm is finally converted into a quadratic optimization problem under the restriction of linear conditions, theoretically, a global optimum point is obtained, and the problem of local extremum which cannot be avoided in a neural network method is solved;
(3) the algorithm converts the actual problem into a high-dimensional feature space through nonlinear transformation, and constructs a linear discriminant function in the high-dimensional space to realize the nonlinear discriminant function in the original space, so that the special property can ensure that the machine has better popularization capability, and simultaneously the problem of dimension is solved skillfully, and the complexity of the algorithm is independent of the dimension of the feature space.
The K-nearest neighbor is the K samples most similar to the samples to be classified, and the class value of the samples to be classified is judged according to the classes of the K samples. In the K-nearest neighbor classifier, an important parameter is the selection of a K value, which is too small to fully reflect the characteristics of a sample to be classified, and if the K value is too large, some samples that are actually dissimilar to the sample to be classified are also included, resulting in an increase in noise and a reduction in classification effect.
The Knn algorithm is defined as: defining the discriminant function as &'s &' si(x)＝kiI ═ 1,2, c, where ki denotes the number of samples belonging to the ω i class among the k nearest neighbors, and the decision rule is that if gi (x) ═ maxki, x ∈ ω i.
The K-nearest neighbor algorithm is a predictive classification algorithm (supervised learning). It does not actually need to generate additional data to describe the rule, and its rule itself is the data (sample). KNN belongs to sample-based learning of machine learning, and is different from inductive learning in that the problem is solved by directly using the existing sample instead of by rule derivation. It does not require data consistency issues, i.e. noise can be present and the sample modification is local and does not require reorganization.
The K-nearest neighbor algorithm synthesizes the categories of K adjacent samples nearest to the unknown sample to predict the category of the unknown sample, and calculates the distance from the unknown sample according to a certain distance formula when selecting the sample to determine whether to select. Its advantages are simple method, stable algorithm and high robustness. The disadvantage is that a large number of samples are required to ensure the accuracy of the data, and furthermore, more importantly, the distance between a large number of samples needs to be calculated, which causes inconvenience in use.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.
Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.
Claims (5)
1. A real-time parathyroid gland recognition method in operation based on near-infrared autofluorescence includes the following steps:
s1, imaging the suspicious parathyroid gland by adopting parathyroid gland imaging equipment;
s2, obtaining fluorescence spectrum data of the suspicious parathyroid gland and surrounding tissues by adopting parathyroid gland spectrum distinguishing equipment;
s3, obtaining a spectrogram and data by giving light with the power of 80mW for 1 second, wherein the test sequence is as follows: firstly, fat and normal thyroid tissue are tested as references, the two tissues can be confirmed by naked eyes in 100 percent, then suspicious parathyroid gland tissues and definite lymph nodes are tested, and parathyroid glands are judged in advance by human beings by taking the standard that the fluorescence peak intensity on a spectrogram exceeds 1 time;
s4, collecting spectrogram data of parathyroid gland, carrying out big data analysis, and forming an artificial intelligence model;
s5, training an intelligent model to form parathyroid gland distinguishing equipment, inputting the imaging atlas of the suspicious parathyroid gland into the parathyroid gland distinguishing equipment, and distinguishing whether the parathyroid gland is the parathyroid gland;
in the step S1, the parathyroid gland imaging device comprises a 785nm laser surface light source, a 3CCD high-sensitivity imaging system and a filtering system, and the filtering system comprises a 808nm long-pass filter and a 822nm narrow-band filter;
in S2, the parathyroid gland spectrum distinguishing device includes a spectrometer and a 785nm laser point light source.
2. The intraoperative real-time parathyroid gland identification method based on near-infrared autofluorescence according to claim 1, characterized in that: the parathyroid gland spectrum distinguishing device in the S2 comprises the following specific steps:
s2.1, obtaining fluorescence spectrograms of suspicious tissues, thyroid glands, fat and lymph nodes;
and S2.2, judging whether the parathyroid gland of the suspicious tissue is determined according to the difference of fluorescence peak intensities by taking the fluorescence intensities of the fat and the normal thyroid gland as a reference.
3. The intraoperative real-time parathyroid gland identification method based on near-infrared autofluorescence according to claim 1, characterized in that: in the S4, the big data analysis method selects Fisher algorithm, Svm algorithm or Knn algorithm.
4. The intraoperative real-time parathyroid gland identification method based on near-infrared autofluorescence according to claim 3, characterized in that: the Fisher algorithm adopts linear discriminant analysis, and considers the inter-class information of the training sample on the basis of using a PCA method to reduce the dimension.
5. The intraoperative real-time parathyroid gland identification method based on near-infrared autofluorescence according to claim 3, characterized in that: the Knn algorithm is defined as: defining a discriminant function as &i(x)＝kiI is 1,2 …, c, where ki represents the number of samples belonging to ω i class in k nearest neighbors, and the decision rule is that if gi (x) maxki, x ∈ ω i.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911016165.7A CN110693458B (en) | 2019-10-24 | 2019-10-24 | Intraoperative real-time parathyroid gland identification method based on near-infrared autofluorescence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911016165.7A CN110693458B (en) | 2019-10-24 | 2019-10-24 | Intraoperative real-time parathyroid gland identification method based on near-infrared autofluorescence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110693458A CN110693458A (en) | 2020-01-17 |
| CN110693458B true CN110693458B (en) | 2022-05-27 |
Family
ID=69201092
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911016165.7A Active CN110693458B (en) | 2019-10-24 | 2019-10-24 | Intraoperative real-time parathyroid gland identification method based on near-infrared autofluorescence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110693458B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20210127957A1 (en) * | 2019-10-31 | 2021-05-06 | Inthesmart Co., Ltd. | Apparatus for intraoperative identification and viability assessment of tissue and method using the same |
| CN114120038A (en) * | 2021-11-26 | 2022-03-01 | 中国科学院长春光学精密机械与物理研究所 | Parathyroid gland identification method based on hyperspectral imaging technology and model training |
| CN116369959B (en) * | 2023-06-05 | 2023-08-11 | 杭州医策科技有限公司 | Parathyroid preoperative positioning method and device based on bimodal CT |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107361744A (en) * | 2017-07-31 | 2017-11-21 | 清华大学 | The identification device and method of a kind of parathyroid gland |
| CN109781701A (en) * | 2019-01-18 | 2019-05-21 | 拉曼兄弟(深圳)科技发展有限公司 | Real-time detection method in a kind of parathyroidectomy based on Raman spectroscopy |
-
2019
- 2019-10-24 CN CN201911016165.7A patent/CN110693458B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107361744A (en) * | 2017-07-31 | 2017-11-21 | 清华大学 | The identification device and method of a kind of parathyroid gland |
| CN109781701A (en) * | 2019-01-18 | 2019-05-21 | 拉曼兄弟(深圳)科技发展有限公司 | Real-time detection method in a kind of parathyroidectomy based on Raman spectroscopy |
Non-Patent Citations (1)
| Title |
|---|
| Automatic Identification of Parathyroid in Optical Coherence Tomography Images;Fang hou 等;《Lasers in Surgery and Medicine》;20171231;第1-7页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110693458A (en) | 2020-01-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110693458B (en) | Intraoperative real-time parathyroid gland identification method based on near-infrared autofluorescence | |
| Adegun et al. | Deep learning-based system for automatic melanoma detection | |
| Ramlakhan et al. | A mobile automated skin lesion classification system | |
| Bird et al. | Infrared spectral histopathology (SHP): a novel diagnostic tool for the accurate classification of lung cancer | |
| US20180025210A1 (en) | Method and system for analyzing biological specimens by spectral imaging | |
| JP5292821B2 (en) | Vein image acquisition device and vein image acquisition method | |
| JP5733032B2 (en) | Image processing apparatus and method, image processing system, program, and recording medium | |
| US20240074664A1 (en) | Apparatus and method for image-guided interventions with hyperspectral imaging | |
| EP3413790A1 (en) | Systems and methods for evaluating pigmented tissue lesions | |
| WO2019130924A1 (en) | Image processing device, endoscope system, image processing method, and program | |
| JP7480708B2 (en) | Information processing device, information processing method, and program | |
| KR102162683B1 (en) | Reading aid using atypical skin disease image data | |
| US10921252B2 (en) | Image processing apparatus and method of operating image processing apparatus | |
| CA2907405A1 (en) | Method and system for analyzing biological specimens by spectral imaging | |
| CN105160289A (en) | Face identification method and terminal | |
| Fernandes et al. | Early skin cancer detection using computer aided diagnosis techniques | |
| Almuntashri et al. | Gleason grade-based automatic classification of prostate cancer pathological images | |
| Ortega et al. | Hyperspectral database of pathological in-vitro human brain samples to detect carcinogenic tissues | |
| Goicoechea-Telleria et al. | Low-cost and efficient hardware solution for presentation attack detection in fingerprint biometrics using special lighting microscopes | |
| JP2006061170A (en) | Discrimination method for skin | |
| Gavrilov et al. | Deep learning based skin lesions diagnosis | |
| Saraswathi et al. | Review on computer aided diagnosis of pancreatic cancer using artificial intelligence system | |
| JP2010181970A (en) | Equipment for biometric identification, biometric identification device, biometric identification system, discrimination reference determination method, biometric identification method, and program | |
| Bedo et al. | Color and texture influence on computer-aided diagnosis of dermatological ulcers | |
| Grechkin et al. | VGG convolutional neural network classification of hyperspectral images of skin neoplasms |
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 |