CN115762756A - Early cognitive disorder scale drawing result judging device - Google Patents
Early cognitive disorder scale drawing result judging device Download PDFInfo
- Publication number
- CN115762756A CN115762756A CN202211382103.XA CN202211382103A CN115762756A CN 115762756 A CN115762756 A CN 115762756A CN 202211382103 A CN202211382103 A CN 202211382103A CN 115762756 A CN115762756 A CN 115762756A
- Authority
- CN
- China
- Prior art keywords
- pic
- hog
- drawing board
- graphic
- algorithm
- 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
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention provides a judging device for an early cognitive disorder scale drawing result, which comprises an inputtable drawing board device, a judging device and a judging device, wherein the inputtable drawing board device is used for acquiring graphic data of a scale drawing question of a tested object drawn on the inputtable drawing board device; the image conversion module is used for obtaining image data by zooming through a zooming algorithm and is electrically connected with equipment capable of inputting a drawing board; and the classifier model is used for judging whether the graphic data scaled by the graphic conversion module is correct or not, and scoring the answer results of the scale drawing questions of the tested object by using '0' or '1', wherein 1 represents correct and 0 represents wrong. The database user data used for training the classifier model is huge and is real user data of hospitals for years, and the adaptability is strong; a paper scale is abandoned, and the judgment result output can be obtained immediately only by reading the input of a new user on the drawing board, so that the whole process only consumes several seconds, and the operation is convenient; the result is stored in the system, the system can be operated in a disconnected network mode, and the personal privacy of the user is guaranteed.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a device for judging a drawing result of an early cognitive impairment scale.
Background
Alzheimer's disease is a common senile neurodegenerative disorder and is also one of the most common types of dementia, and its clinical features include memory deficits, loss of environmental cognition and impairment of motor function, and psychotic symptoms. Mild Cognitive Impairment (MCI) is a transitional stage between natural aging and alzheimer's disease, and primary screening of their related populations is typically paper-based scale screening, for example: a simple mental state scale (MMSE), a montreal cognitive assessment scale (MoCA), or a clinical dementia assessment (CDR), among others.
Due to different cognitive abilities of users, the scales usually have the problem of long time consumption, meanwhile, when professional medical care personnel process the scale results afterwards, the required scoring time is also long, and particularly the accuracy rate is possibly reduced by long-time manual processing operation for distinguishing drawing questions. Therefore, how to design an efficient and convenient scale drawing result determination device becomes an urgent problem to be solved.
Disclosure of Invention
The invention mainly aims to provide an early cognitive disorder scale drawing result judging device which can be used for efficiently and conveniently judging the cognitive level of a user.
In order to realize the purpose, the invention adopts the following technical scheme:
the invention provides a device for judging the drawing result of an early cognitive impairment scale, which comprises:
the system comprises an inputtable drawing board device, a meter drawing item input device and a meter drawing item output device, wherein the inputtable drawing board device is used for acquiring graphic data of the meter drawing item drawn by a tested object on the inputtable drawing board device;
the graphic conversion module is used for zooming graphic data acquired from the inputtable drawing board equipment through a zooming algorithm and is electrically connected with the inputtable drawing board equipment;
and the classifier model is used for judging whether the graphic data scaled by the graphic conversion module is correct or not, and scoring the answer result of the scale drawing questions of the tested object by using '0' or '1', wherein 1 represents correct and 0 represents wrong.
Further, the drawing board input device is a digital drawing board.
Further, the graphical data includes closed and non-closed graphical patterns of the subject drawn on an inputtable sketchpad device over a given time.
Further, the scaling algorithm comprises: extracting the position information of a graphic pattern drawn by the tested object on the equipment capable of inputting the drawing board, and calculating the original size of graphic pattern data; and scaling the size of the graphic pattern data by utilizing a bilinear interpolation algorithm.
The classifier model is trained by positive and negative samples obtained in advance, the positive and negative samples consist of recorded paper scale drawing question answer results of a healthy control group and a mild cognitive impairment patient group, the answer result of the healthy control group corresponds to a positive sample, and the answer result of the mild cognitive impairment patient group corresponds to a negative sample; under the condition that positive and negative samples are unchanged, the classifier model is obtained without secondary training; after the number of the positive samples and the negative samples is changed, the classifier model needs to be retrained to obtain, and the required data synchronously change.
Further, the training module firstly extracts the features of the positive and negative samples, and then classifies the results after feature extraction by using a classifier model.
Further, the algorithm of feature extraction is an improved radial gradient histogram algorithm, and the classifier is a support vector machine or a decision tree.
Furthermore, the feature extraction is to divide the image into a plurality of connected regions, then calculate the edge direction of multi-directional gradient histograms or pixel points in each region, and the gradient histograms or pixel point edge directions of all regions of the image form the feature vector of the image.
Specifically, the picture received by the classifier model is first halved, and the picture before halving is pic 1 The picture after bisection is pic 21 And pic 22 (ii) a Separately aligning pic by HOG algorithm 1 、pic 21 And pic 22 Extracting features to obtain corresponding feature vector of hog 1 、hog 21 And hog 22 Let hog 21 And hog 22 Forming a new characteristic vector hog by connecting the head and the tail 2 (ii) a Then pic 21 Halving to obtain picture pic 31 And pic 32 Then pic 22 Halving to obtain picture pic 33 And pic 34 For picture pic 33 And pic 34 Similarly, feature extraction is carried out in a blocking mode by utilizing an HOG algorithm, and the obtained feature vectors are connected end to form a new feature vector HOG 3 To this end, the first picture pic 1 After the feature extraction is finished, the feature vector is hog 1 、hog 2 And hog 3 A set of (a).
Further, the resolution of the classifier model identification image is variable, preferably 96 pixels by 128 pixels.
The use method of the early cognitive impairment scale drawing result judging device comprises the following steps:
s1, after a tested object reads a scale drawing question, drawing graphic data on an inputtable drawing board device, and transmitting the graphic data drawn by the tested object to a graphic conversion module by the inputtable drawing board device;
s2, after the graph conversion module receives graph data which can be input and transmitted by the drawing board equipment, the graph data is zoomed by utilizing a zooming algorithm to meet the requirement of the classifier model;
and S3, the classifier model gives a final judgment result according to the graphic data scaled by the graphic conversion module, and the answer result of the scale drawing question of the tested object is scored by using 0 or 1, wherein 1 represents correct and 0 represents wrong.
The invention has the beneficial effects that:
the invention provides a device for judging a drawing result of an early cognitive impairment scale. Firstly, after a tested object reads a scale drawing question, acquiring graphic data of the tested object drawn on equipment capable of inputting a drawing board; then, the acquired data is transmitted to a graph conversion module, the graph conversion module utilizes a scaling algorithm to scale the graph data so as to meet the requirement of a classifier model, the scaling algorithm comprises a bilinear inner difference algorithm, the calculation amount of the algorithm is small, and the calculation speed is high; finally, the classifier model gives a final judgment result aiming at the graph input by the graph conversion module; the classifier model is obtained by training pre-obtained positive and negative samples, and under the condition that the positive and negative samples are not changed, the classifier model is obtained without secondary training; the training method of the classifier model comprises the steps of firstly extracting features of positive and negative samples, and then classifying the results after feature extraction by using the classifier model, wherein the algorithm of feature extraction is an improved radial gradient histogram algorithm; the database user data used for training the classifier model is huge and is real user data of hospitals for years, and the adaptability is strong; a paper scale is abandoned, and the judgment result output can be obtained immediately only by reading the input of a new user on the drawing board, so that the whole process only consumes several seconds, and the operation is convenient and fast; the result is stored in the system, the system can be disconnected from the network to operate, and the personal privacy of the user is guaranteed. In conclusion, the invention has the advantages of high efficiency, wide application range and easy operation.
Drawings
FIG. 1 is a schematic flow diagram of the present invention;
FIG. 2 is a diagram of a system scale topic in accordance with the present invention;
FIG. 3 is a schematic diagram of a positive sample of a database according to the present invention;
FIG. 4 is a schematic diagram of a negative sample of a database according to the present invention;
FIG. 5 is a schematic view of a test example of the present invention;
FIG. 6 is a schematic diagram of the source and accuracy of the test object according to the present invention.
Detailed Description
The invention provides a device for judging the drawing result of an early cognitive impairment scale, which comprises:
the system comprises an inputtable drawing board device, a database and a database, wherein the inputtable drawing board device is used for acquiring graphic data of a scale drawing theme drawn by a tested object on the inputtable drawing board device, and the inputtable drawing board device is a digital drawing board;
a graphic conversion module for scaling closed and non-closed graphic patterns drawn on an inputtable sketchpad device within a given time from a subject acquired from the inputtable sketchpad device by a scaling algorithm, the graphic conversion module being electrically connected to the inputtable sketchpad device; the scaling algorithm comprises: extracting the position information of a graphic pattern of a tested object which can be input on drawing board equipment, calculating the original size of graphic pattern data, and scaling the size of the graphic pattern data by utilizing a bilinear interpolation algorithm;
the training module is used for training a classifier model by using pre-obtained positive and negative samples, wherein the positive and negative samples comprise recorded paper scale drawing question answering results of a healthy control group and a mild cognitive impairment patient group, the answer results of the healthy control group correspond to the positive samples, and the answer results of the mild cognitive impairment patient group correspond to the negative samples; the training method of the classifier model comprises the following steps: firstly, extracting the features of positive and negative samples, and then classifying the results after feature extraction by using a classifier model; the feature extraction algorithm is an improved radial gradient histogram algorithm, and the classifier is a support vector machine or a decision tree.
The classifier model is used for judging whether the graphic data zoomed by the graphic conversion module is correct or not, and scoring the scale drawing question result of the tested object by using '0' or '1', wherein 1 represents correct and 0 represents error;
as shown in fig. 1, the procedure of determining the cognitive state of the patient using the early cognitive impairment scale mapping result determination device is as follows:
s1, selecting healthy people and mild disorder patients in the community elderly center and hospital outpatient clinic range for the tested object, wherein the drawing title of the scale is the 19 th title (figure 2) of a simple Mental State scale (MMSE), and connecting a digital drawing board with a computer through a USB interface and using the digital drawing board in combination with related software on the computer.
MMSE is displayed on a computer screen, a tested object operates a painting brush to draw on a digital drawing board according to the requirement of a theme, and the digital drawing board records the drawn graph. Setting four points of the graph edge in clockwise direction as the horizontal and vertical axes as the leftmost, the uppermost, the rightmost and the lowermost pointsABCD, coordinates of which are: a (x) a ,y a )、B(x b ,y b )、C(x c ,y c ) And D (x) d ,y d ) The drawn pattern is usually irregular and has a large number of blank areas around it, which do not belong to the object to be processed and increase the amount of calculation, so that the blank areas are removed while the smallest rectangular area containing the drawing pattern of the object to be tested is retained. The positions of the four edges of the rectangle are determined by the ABCD points, and the ordinate value of the upper edge of the rectangle is y b -1, lower edge ordinate value y d +1, x is the horizontal coordinate value of the left edge a -1, right edge abscissa value x c +1, the rectangular area size is (y) d -y b +2)*(x c -x a + 2), and the graphics data of the rectangular area is transmitted to the graphics conversion module.
And S2, the graphic conversion module receives the graphic data of the rectangular area in the step S1, and the sizes of the graphic data of the rectangular area are different due to different tested objects. Therefore, according to the picture size required by the classifier model as the scaling standard, in the present embodiment, the resolution size of the image recognized by the classifier model is 96 pixels by 128 pixels, where 96 is the picture width and 128 is the picture length. The adopted scaling algorithm is a bilinear interpolation value algorithm, the algorithm multiplies the obtained pixel information of four pixel points of a point near an original sampling point (floating point number) by weight to obtain pixel information of a new image, interpolation is carried out in the horizontal direction or the vertical direction at will, and the sequence of the interpolation direction does not influence the final result. The algorithm has the advantages of small calculation amount and high operation speed. The image interpolated by the method has better continuity and high image quality, overcomes the defect of discontinuous gray values and can effectively resist the sawtooth effect of the image. And scaling the size of the graphic data by utilizing a bilinear interpolation algorithm, wherein the scaled graphic size is 96 pixels by 128 pixels.
S3, providing a final judgment result of the graph data input by the graph conversion module by using a classifier model, wherein the classifier model is obtained by training a pre-obtained positive sample and a pre-obtained negative sample; the positive and negative samples are composed of paper scale drawing question answer results of a health control group and a mild cognitive impairment patient group, the answer results of the health control group correspond to positive samples, the answer results of the mild cognitive impairment patient group correspond to negative samples, as shown in fig. 3-4, the positive samples are correct drawing answer sets screened out after manual scoring by a hospital doctor, and the negative samples are corresponding wrong drawing answer sets.
The training method of the classifier model comprises the steps of carrying out feature extraction on positive and negative samples and classifying results by utilizing the classifier model. Firstly, creating a positive sample folder and a negative sample folder, and putting the positive sample data and the negative sample data into the positive sample folder and the negative sample folder; and then, performing feature extraction on all data in the sample folder by using a modified Histogram of Oriented Gradients (HOG) algorithm to obtain feature sets pos and neg corresponding to the positive sample and the negative sample respectively. The core idea of the algorithm is to describe the whole object by using the distribution of the light intensity or edge direction of the surface of the object. The specific operation is that the image is divided into a plurality of connection areas, then multidirectional gradient histograms or edge directions of pixel points are calculated in each area, the gradient histograms or the edge directions of the pixel points of all the areas of the image form a descriptor of the image, and the descriptor can be used as a feature vector of the image; firstly, dividing a picture received by a classifier model into two halves, and marking the picture before the two halves as pic 1 The picture after bisection is denoted pic 21 And pic 22 The area of the last two pixels is half of the area of the last two pixels, namely 96 pixels by 64 pixels; separately aligning pic by HOG algorithm 1 、pic 21 And pic 22 Extracting features, and recording the obtained corresponding feature vector as hog 1 、hog 21 And hog 22 G. will hog 21 And hog 22 The head and the tail are connected to form a new characteristic vector which is marked as hog 2 (ii) a Then pic 21 Halving to obtain picture pic 31 And pic 32 Then mix pic with 22 Halving to obtain picture pic 33 And pic 34 Then pic 31 、pic 32 、pic 33 And pic 34 The area size is 48 pixels by 64 pixels, and similarly, the block division is carried out by using the HOG algorithmExtracting features, connecting the obtained feature vectors end to form a new feature vector, and marking as hog 3 . To this end, the present invention deals with the first picture pic 1 After the feature extraction is finished, the feature vector is hog 1 、hog 2 And hog 3 A set of (a); and then, training a classifier model by using a support vector machine binary classification algorithm, wherein the radial basis kernel function is adopted for training.
The classifier model is obtained without secondary training under the condition that positive and negative sample data are not changed. If a new user database is added, the model can be optimized or iterated, that is, the number of positive and negative samples can be increased or reduced, and meanwhile, the classifier model is obtained by retraining correspondingly, and the required value is changed correspondingly.
After the classifier model is trained, the data which is input by a user and then is zoomed by the graph conversion module is judged, the display form of the judgment result is score, namely 1 or 0, 1 represents that the graph drawn by the user is in accordance with the requirement of the question and is judged to be correct, 0 represents that the graph drawn by the user is not in accordance with the requirement of the question and is judged to be wrong, and the judgment result is stored in the system. FIG. 5 is a schematic diagram of a test example of the present invention. FIG. 6 is a schematic diagram of the sources and accuracy of the test objects of the present invention, and as can be seen from FIG. 6, the main test objects are divided into two categories: the senior community center and the hospital outpatient service comprise healthy senior citizens and senior patients diagnosed as mild cognitive impairment by the hospital, and four groups of people are provided. The histogram result shows that the accuracy of the judging device is more than 90%, and the judging effect is good.
In the invention, a user database for training a classifier model is huge and is real user data of a hospital for years; the operation is convenient and fast, a paper scale is abandoned, the output of a judgment result can be obtained immediately only by reading the input of a new user on the drawing board, and the whole process only takes several seconds; the result is stored in the system, and the personal privacy of the user is guaranteed.
The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above-mentioned embodiments, and all technical solutions belonging to the idea of the present invention belong to the protection scope of the present invention. It should be noted that modifications and embellishments within the scope of the invention may be made by those skilled in the art without departing from the principle of the invention.
Claims (10)
1. An early cognitive impairment scale drawing result discrimination device, comprising:
the system comprises an inputtable drawing board device, a meter drawing item input device and a meter drawing item output device, wherein the inputtable drawing board device is used for acquiring graphic data of the meter drawing item drawn by a tested object on the inputtable drawing board device;
the graphic conversion module is used for zooming the graphic data acquired from the equipment capable of inputting the drawing board through a zooming algorithm and is electrically connected with the equipment capable of inputting the drawing board;
and the classifier model is used for judging whether the graphic data scaled by the graphic conversion module is correct or not, and scoring the answer results of the scale drawing questions of the tested object by using '0' or '1', wherein 1 represents correct and 0 represents wrong.
2. The device for distinguishing the drawing result of the early cognitive impairment scale as claimed in claim 1, wherein the inputtable drawing board device is a digital drawing board.
3. The device for distinguishing the drawing result of the early cognitive impairment scale as claimed in claim 1, wherein the graphic data comprises closed and non-closed graphic patterns drawn by the subject on a drawing board input device within a given time.
4. The device for distinguishing the drawing result of the early cognitive impairment scale as claimed in claim 1, wherein the scaling algorithm comprises: extracting the position information of a graphic pattern drawn by the tested object on the equipment capable of inputting the drawing board, and calculating the original size of graphic pattern data; and scaling the size of the graphic pattern data by utilizing a bilinear interpolation algorithm.
5. The device as claimed in claim 1, further comprising a training module for training the classifier model by pre-obtained positive and negative samples, wherein the positive and negative samples comprise the answer results of the paper scale drawing questions of the health control group and the mild cognitive impairment patient group, the answer results of the health control group correspond to the positive samples, and the answer results of the mild cognitive impairment patient group correspond to the negative samples.
6. The device for distinguishing the drawing result of the early cognitive impairment scale according to claim 5, wherein the training module performs feature extraction on the positive and negative samples and then classifies the result after feature extraction by using a classifier model.
7. The device for distinguishing the drawing result of the early cognitive impairment scale according to claim 6, wherein the feature extraction algorithm is an improved radial gradient histogram algorithm, and the classifier is a support vector machine or a decision tree.
8. The device as claimed in claim 7, wherein the feature extraction is to divide the image into a plurality of connected regions, and then calculate the edge directions of multi-directional gradient histograms or pixels in each region, and the gradient histograms or pixel edge directions of all regions of the image constitute the feature vector of the image.
9. The device for distinguishing the drawing result of the early cognitive impairment scale according to claim 8, wherein the feature extraction method is specifically as follows:
firstly, dividing a picture received by a classifier model into two halves, wherein the picture before the two halves is pic 1 The picture after halving is pic 21 And pic 22 (ii) a Separately aligning pic by HOG algorithm 1 、pic 21 And pic 22 Extracting features to obtain corresponding feature vector of hog 1 、hog 21 And hog 22 Let hog 21 And hog 22 Form a new characteristic vector hog by connecting head and tail 2 (ii) a Then pic 21 Halving to obtain picture pic 31 And pic 32 Then mix pic with 22 Halving to obtain picture pic 33 And pic 34 For picture pic 33 And pic 34 And carrying out feature extraction by utilizing the HOG algorithm in a blocking manner, and connecting the obtained feature vectors end to form a new feature vector HOG 3 To this end, the first picture pic 1 After the feature extraction is finished, the feature vector is hog 1 、hog 2 And hog 3 A collection of (a).
10. The device according to claim 5, wherein the classifier model identifies the resolution of the image as 96 pixels by 128 pixels.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211382103.XA CN115762756A (en) | 2022-11-07 | 2022-11-07 | Early cognitive disorder scale drawing result judging device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211382103.XA CN115762756A (en) | 2022-11-07 | 2022-11-07 | Early cognitive disorder scale drawing result judging device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115762756A true CN115762756A (en) | 2023-03-07 |
Family
ID=85356735
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211382103.XA Pending CN115762756A (en) | 2022-11-07 | 2022-11-07 | Early cognitive disorder scale drawing result judging device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115762756A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116883794A (en) * | 2023-09-07 | 2023-10-13 | 北京智精灵科技有限公司 | Spatial structure cognitive ability evaluation method and system based on graph attention network |
| CN116883794B (en) * | 2023-09-07 | 2024-05-31 | 北京智精灵科技有限公司 | Spatial structure cognitive ability evaluation method and system based on graph attention network |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109409329A (en) * | 2018-11-14 | 2019-03-01 | 科大讯飞股份有限公司 | Method of discrimination, appraisal procedure, device, equipment and the storage medium of figure topic |
| CN111081373A (en) * | 2019-12-16 | 2020-04-28 | 中国科学院深圳先进技术研究院 | Cognitive disorder training and evaluating system and method |
| CN112241691A (en) * | 2020-09-25 | 2021-01-19 | 天津大学 | Channel ice condition intelligent identification method based on unmanned aerial vehicle inspection and image characteristics |
| US20210153801A1 (en) * | 2019-11-26 | 2021-05-27 | The Chinese University Of Hong Kong | Methods based on an analysis of drawing behavior changes for cognitive dysfunction screening |
| CN114334123A (en) * | 2020-10-10 | 2022-04-12 | 北京师范大学 | Cognition assessment system suitable for mild cognitive impairment rapid detection |
-
2022
- 2022-11-07 CN CN202211382103.XA patent/CN115762756A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109409329A (en) * | 2018-11-14 | 2019-03-01 | 科大讯飞股份有限公司 | Method of discrimination, appraisal procedure, device, equipment and the storage medium of figure topic |
| US20210153801A1 (en) * | 2019-11-26 | 2021-05-27 | The Chinese University Of Hong Kong | Methods based on an analysis of drawing behavior changes for cognitive dysfunction screening |
| CN111081373A (en) * | 2019-12-16 | 2020-04-28 | 中国科学院深圳先进技术研究院 | Cognitive disorder training and evaluating system and method |
| CN112241691A (en) * | 2020-09-25 | 2021-01-19 | 天津大学 | Channel ice condition intelligent identification method based on unmanned aerial vehicle inspection and image characteristics |
| CN114334123A (en) * | 2020-10-10 | 2022-04-12 | 北京师范大学 | Cognition assessment system suitable for mild cognitive impairment rapid detection |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116883794A (en) * | 2023-09-07 | 2023-10-13 | 北京智精灵科技有限公司 | Spatial structure cognitive ability evaluation method and system based on graph attention network |
| CN116883794B (en) * | 2023-09-07 | 2024-05-31 | 北京智精灵科技有限公司 | Spatial structure cognitive ability evaluation method and system based on graph attention network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109389129B (en) | Image processing method, electronic device and storage medium | |
| CN103518224B (en) | Method for analysing microbial growth | |
| WO2022001571A1 (en) | Computing method based on super-pixel image similarity | |
| CN111986150B (en) | The method comprises the following steps of: digital number pathological image Interactive annotation refining method | |
| CN108596046A (en) | A kind of cell detection method of counting and system based on deep learning | |
| Zhang et al. | Automated semantic segmentation of red blood cells for sickle cell disease | |
| Watson | Perimetric complexity of binary digital images: Notes on calculation and relation to visual complexity | |
| CN110503054B (en) | Text image processing method and device | |
| CN107665492A (en) | Colon and rectum panorama numeral pathological image tissue segmentation methods based on depth network | |
| CN104484886B (en) | A kind of dividing method and device of MR images | |
| CN109740572A (en) | A kind of human face in-vivo detection method based on partial color textural characteristics | |
| CN107909102A (en) | A kind of sorting technique of histopathology image | |
| CN111160194B (en) | Static gesture image recognition method based on multi-feature fusion | |
| CN108305253A (en) | A kind of pathology full slice diagnostic method based on more multiplying power deep learnings | |
| CN115546605A (en) | Training method and device based on image labeling and segmentation model | |
| CN110728666A (en) | Typing method and system for chronic nasosinusitis based on digital pathological slide | |
| CN113160185A (en) | Method for guiding cervical cell segmentation by using generated boundary position | |
| CN115909006A (en) | Mammary tissue image classification method and system based on convolution Transformer | |
| CN103377379A (en) | Text detection device and method and text information extraction system and method | |
| CN110298236A (en) | A kind of braille automatic distinguishing method for image and system based on deep learning | |
| CN115762756A (en) | Early cognitive disorder scale drawing result judging device | |
| CN113989269B (en) | Traditional Chinese medicine tongue image tooth trace automatic detection method based on convolutional neural network multi-scale feature fusion | |
| Lakshmi et al. | Real-Time Hand Gesture Recognition for Improved Communication with Deaf and Hard of Hearing Individuals | |
| Kumar et al. | An effective age classification using topological features based on compressed and reduced grey level model of the facial skin | |
| CN113222887A (en) | Deep learning-based nano-iron labeled neural stem cell tracing method |
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 |