CN104537379A - High-precision automatic tongue partition method - Google Patents
High-precision automatic tongue partition method Download PDFInfo
- Publication number
- CN104537379A CN104537379A CN201410823935.XA CN201410823935A CN104537379A CN 104537379 A CN104537379 A CN 104537379A CN 201410823935 A CN201410823935 A CN 201410823935A CN 104537379 A CN104537379 A CN 104537379A
- Authority
- CN
- China
- Prior art keywords
- pixel
- value
- tongue
- color
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20112—Image segmentation details
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a high-precision automatic tongue partition method. The high-precision automatic tongue partition method comprises the following steps that (1) a cascade classifier used for detecting the tongue is trained; (2) an image is input, the tongue is roughly positioned through the cascade classifier, and an initial color sampling area is determined; (3) the accurate position of the tongue is obtained through an iterative histogram projection method, and points in the image are divided into three types, namely foreground points, background points and points to the recognized; (4) the accurate partition of the tongue is worked out through the mode of establishing a local linear estimation model for an area to be recognized and conducting solving through a global restriction equation. The method is achieved in a full automatic mode, and the tongue can be partitioned as long as the image is input; the partition speed is high, the whole process of detecting, positioning and partitioning can be completed by 3-4 seconds, and efficiency is ten times or more that of manual partition; the partition precision is high, local studying and global restriction solving are conducted through pixels, the finial partition effect is corrected to the pixel level, and the effect is far better than that of a traditional partition method.
Description
Technical field
The present invention relates to a kind of high-precision tongue body automatic division method, belong to Pattern recognition and image processing field.
Background technology
The basic skills of tcm diagnosis has four kinds, that is: hope, hear, ask, cut.An important component part in observation is exactly tongue picture diagnosis.Because lingual diagnosis has noncontact, the painless and advantage such as to have no side effect, this makes it meet very much modern medicine theory, is easily applied in the emerging medical skills such as machine medical assistance, remote medical consultation with specialists simultaneously.Such as, but there is certain subjectivity in tongue picture diagnosis, and is easily subject to the impact of objective environment, the factors such as illumination.The tongue picture diagnosis that machine is assisted solves the problems referred to above to a certain extent, has promoted development and the internationalization of traditional tongue picture diagnosis.Tongue body segmentation is the steps necessary that machine assists tongue picture to diagnose also is committed step, and the quality of its segmentation effect directly affects the accuracy of diagnostic result.The tongue body segmentation of the current overwhelming majority all by completing by hand, the bad and inefficiency of segmentation effect.And the existing automatic division method shortcoming that all existence and stability is poor, be difficult to the needs meeting actual diagnosis.
Summary of the invention
The object of the invention is to the deficiency existed for prior art, propose a kind of high-precision tongue body automatic division method.The method adopts histogram projection to carry out tongue location, obtains the foreground point and background dot determined, setting up local linear estimation model, utilizing the mode of global restriction equation solution to calculate the Accurate Segmentation of tongue by treating identified region.The present invention effectively overcomes large, the consuming time length of workload of conventional segmentation tongue body, the shortcoming of segmentation precision difference.
In order to achieve the above object, the present invention adopts following technical proposals:
A kind of high-precision tongue body automatic division method, comprises the steps:
(1) one is trained for detecting the cascade classifier of tongue;
(2) input a pictures, carried out the coarse localization of tongue body by cascade classifier, determine priming color sample area;
(3) use the histogram projection method of iteration, obtain the exact position of tongue, image mid point is divided three classes: foreground point, background dot and point to be identified;
(4) solve image configuration global restriction equation, obtain all prospects, namely tongue body pixel, completes segmentation.
The concrete steps of above-mentioned steps (1) are as follows:
(11) use same money digital camera, under constant light source, gather a collection of containing tongued medical;
(12) photo photographed is screened, manual mark tongue region;
(13) one is trained to detect cascade classifier for tongue.
The concrete steps of above-mentioned steps (2) are as follows:
(21) employing and the middle the same terms of step (11) take a width tongue photo;
(22) cascade classifier utilizing training to obtain searches the tongue position in photo, and uses rectangle frame
rmark;
(23) geometric center and rectangle frame is determined
ridentical, the length of side is its little rectangle of 1/4
r, as color samples region.
The concrete steps of above-mentioned steps (3) are as follows:
(31) color reduction operation is carried out to picture, make its three Color Channels all be reduced to 32 kinds of color depths, namely in its rgb space, retain 32*32*32=32768 kind color;
(32) to color samples region
rcarry out Color Statistical, obtain its color histogram and by its normalization;
(33) by the normalized histogram projection of the picture after color reduction, that is: travel through all pixels of original image, to each pixel, record the color value under its RGB pattern
c; Search normalized histogram, obtain current color value
ccorresponding histogram value
v, and as the value of current pixel; After operation terminates, obtain the single channel image that a width is new, each pixel value represents it and belongs to prospect, namely belongs to the probability of tongue;
(341) image that step (33) obtains is carried out thresholding process, threshold value gets 0.05, generates a width bianry image;
(342) carry out gaussian filtering to bianry image, discrete point is wherein polymerized further, and hole is filled; Now there is multiple connected region in image;
(343) extract the largest connected region of image, its boundary rectangle is designated as
r; Calculate, with it, there is identical geometric center, and the length of side is the little rectangle of its half
r new if, little rectangle
r new with old sample area
rcompare, position and area discrepancy are comparatively large, then by little rectangle
r new assignment is given
r, be transferred to step (32) as new sample area.Otherwise by little rectangle
r new the point surrounded is labeled as foreground point, by large rectangle
rpoint is in addition labeled as background dot, and other is labeled as pixel to be determined.
The concrete steps of above-mentioned steps (4) are as follows:
(41) travel through all pixels to be identified, to each point, utilize pixel in the neighborhood of place 3*3 to set up linear color estimation model to it:
(411) the corresponding weight vector of pixel is solved:
In above formula, the pixel number of image is designated as
n, for label be
ipixel,
imeet
1<i<n.Record its color value
.The color matrix recording all pixels in its neighborhood is
.
ifor unit matrix, coefficient
tvalue is 0.000001,
f i for required weight vector; Namely for pixel
i, its pixel color value can use weight vector
f i with its neighborhood territory pixel color value Linearly Representation;
(412) weight vector expansion: namely for pixel
i, its pixel color value can use weight vector
ξ i with pixel color value Linearly Representation in all images:
If wherein pixel
jfor pixel
ineighbor, its value is weight vector
f i middle respective value, otherwise be 0;
(42) image configuration global restriction equation is solved, obtain all foreground pixels point, complete segmentation;
(421) image configuration global restriction equation is solved, obtains foreground area masking-out, value be 1 point be foreground point, value be-1 point be background dot:
In above formula, weight matrix
, size is
n*n;
cfor diagonal matrix, if pixel
jfor marked pixels, then diagonal matrix
c?
jrow diagonal element value is constant 100000, otherwise value is 0;
α * for the classified information vector of marker image vegetarian refreshments, foreground point value 1, background dot value-1;
(422) will
αmiddle value is the pixel extraction of 1 and preserves into the new image of a width according to its former coordinate, completes tongue body segmentation.
The method of high-precision tongue body automatic division method of the present invention and traditional-handwork segmentation tongue body is than advantage:
The method is full-automatic realization, easy to use, only needs to input the segmentation that namely picture can complete tongue body; Splitting speed is fast, only needs the overall process that namely can complete detection, location 3 ~ 4 seconds and split, and efficiency is 10 times of manual segmentation; Segmentation precision is high, by pixel carry out local study and global restriction solve, make final segmentation effect be accurate to pixel scale, effect is much better than manual segmentation.
Accompanying drawing explanation
Fig. 1 is the overall block flow diagram of high-precision tongue body automatic division method of the present invention.
Fig. 2 is high-precision tongue body automatic division method one width lingual surface image of the present invention.
Embodiment
Below in conjunction with accompanying drawing, embodiments of the invention are described in further detail.
As shown in Figure 1, high-precision tongue body automatic division method of the present invention, its concrete steps are as follows:
A kind of high-precision tongue body automatic division method, comprises the steps:
(1) one is trained for detecting the cascade classifier of tongue;
(2) input a pictures, carried out the coarse localization of tongue body by cascade classifier, determine priming color sample area;
(3) use the histogram projection method of iteration, obtain the exact position of tongue, image mid point is divided three classes: foreground point, background dot and point to be identified;
(4) solve image configuration global restriction equation, obtain all prospects, namely tongue body pixel, completes segmentation.
The concrete steps of described step (1) are as follows:
(11) use same money digital camera, under constant light source, gather a collection of containing tongued medical;
(12) photo photographed is screened, manual mark tongue region;
(13) one is trained to detect cascade classifier for tongue.
The concrete steps of described step (2) are as follows:
(21) employing and the middle the same terms of step (11) take a width tongue photo;
(22) cascade classifier utilizing training to obtain searches the tongue position in photo, and uses rectangle frame
rmark;
(23) geometric center and rectangle frame is determined
ridentical, the length of side is its little rectangle of 1/4
r, as color samples region.
The concrete steps of described step (3) are as follows:
(31) color reduction operation is carried out to picture, make its three Color Channels all be reduced to 32 kinds of color depths, namely in its rgb space, retain 32*32*32=32768 kind color;
(32) to color samples region
rcarry out Color Statistical, obtain its color histogram and by its normalization;
(33) by the normalized histogram projection of the picture after color reduction, that is: travel through all pixels of original image, to each pixel, record the color value under its RGB pattern
c; Search normalized histogram, obtain current color value
ccorresponding histogram value
v, and as the value of current pixel; After operation terminates, obtain the single channel image that a width is new, each pixel value represents it and belongs to prospect, namely belongs to the probability of tongue;
(341) image that step (33) obtains is carried out thresholding process, threshold value gets 0.05, generates a width bianry image;
(342) carry out gaussian filtering to bianry image, discrete point is wherein polymerized further, and hole is filled; Now there is multiple connected region in image;
(343) extract the largest connected region of image, its boundary rectangle is designated as
r; Calculate, with it, there is identical geometric center, and the length of side is the little rectangle of its half
r new if, little rectangle
r new with old sample area
rcompare, position and area discrepancy are comparatively large, then by little rectangle
r new assignment is given
r, be transferred to step (32) as new sample area.Otherwise by little rectangle
r new the point surrounded is labeled as foreground point, by large rectangle
rpoint is in addition labeled as background dot, and other is labeled as pixel to be determined.
The concrete steps of described step (4) are as follows:
(41) travel through all pixels to be identified, to each point, utilize pixel in the neighborhood of place 3*3 to set up linear color estimation model to it:
(411) the corresponding weight vector of pixel is solved:
In above formula, the pixel number of image is designated as
n, for label be
ipixel,
imeet
1<i<n.Record its color value
.The color matrix recording all pixels in its neighborhood is
.
ifor unit matrix, coefficient
tvalue is 0.000001,
f i for required weight vector; Namely for pixel
i, its pixel color value can use weight vector
f i with its neighborhood territory pixel color value Linearly Representation;
(412) weight vector expansion: namely for pixel
i, its pixel color value can use weight vector
ξ i with pixel color value Linearly Representation in all images:
If wherein pixel
jfor pixel
ineighbor, its value is weight vector
f i middle respective value, otherwise be 0;
(42) image configuration global restriction equation is solved, obtain all foreground pixels point, complete segmentation;
(421) image configuration global restriction equation is solved, obtains foreground area masking-out, value be 1 point be foreground point, value be-1 point be background dot:
In above formula, weight matrix
, size is
n*n;
cfor diagonal matrix, if pixel
jfor marked pixels, then diagonal matrix
c?
jrow diagonal element value is constant 100000, otherwise value is 0;
α * for the classified information vector of marker image vegetarian refreshments, foreground point value 1, background dot value-1;
(422) will
αmiddle value is the pixel extraction of 1 and preserves into the new image of a width according to its former coordinate, completes tongue body segmentation.
In the present embodiment, 480 width are as shown in the table containing the segmentation result of tongued medical image, and final tongue body segmentation result is the ratio that accuracy rate is higher is 95.69%, and the ratio that recall rate is higher is 92.15%.
Table 1 480 width contains the segmentation result of tongue medical image
Claims (5)
1. a high-precision tongue body automatic division method, is characterized in that, comprise the steps:
(1) one is trained for detecting the cascade classifier of tongue;
(2) input a pictures, carried out the coarse localization of tongue body by cascade classifier, determine priming color sample area;
(3) use the histogram projection method of iteration, obtain the exact position of tongue, image mid point is divided three classes: foreground point, background dot and point to be identified;
(4) solve image configuration global restriction equation, obtain all prospects, namely tongue body pixel, completes segmentation.
2. high-precision tongue body automatic division method according to claim 1, is characterized in that, the concrete steps of described step (1) are as follows:
(11) use same money digital camera, under constant light source, gather a collection of containing tongued medical;
(12) photo photographed is screened, manual mark tongue region;
(13) one is trained to detect cascade classifier for tongue.
3. high-precision tongue body automatic division method according to claim 1, is characterized in that, the concrete steps of described step (2) are as follows:
(21) employing and the middle the same terms of step (11) take a width tongue photo;
(22) cascade classifier utilizing training to obtain searches the tongue position in photo, and uses rectangle frame
rmark;
(23) geometric center and rectangle frame is determined
ridentical, the length of side is its little rectangle of 1/4
r, as color samples region.
4. high-precision tongue body automatic division method according to claim 1, is characterized in that, the concrete steps of described step (3) are as follows:
(31) color reduction operation is carried out to picture, make its three Color Channels all be reduced to 32 kinds of color depths, namely in its rgb space, retain 32*32*32=32768 kind color;
(32) to color samples region
rcarry out Color Statistical, obtain its color histogram and by its normalization;
(33) by the normalized histogram projection of the picture after color reduction, that is: travel through all pixels of original image, to each pixel, record the color value under its RGB pattern
c; Search normalized histogram, obtain current color value
ccorresponding histogram value
v, and as the value of current pixel; After operation terminates, obtain the single channel image that a width is new, each pixel value represents it and belongs to prospect, namely belongs to the probability of tongue;
(341) image that step (33) obtains is carried out thresholding process, threshold value gets 0.05, generates a width bianry image;
(342) carry out gaussian filtering to bianry image, discrete point is wherein polymerized further, and hole is filled; Now there is multiple connected region in image;
(343) extract the largest connected region of image, its boundary rectangle is designated as
r; Calculate, with it, there is identical geometric center, and the length of side is the little rectangle of its half
r new if, little rectangle
r new with old sample area
rcompare, position and area discrepancy are comparatively large, then by little rectangle
r new assignment is given
r, be transferred to step (32) as new sample area, otherwise by little rectangle
r new the point surrounded is labeled as foreground point, by large rectangle
rpoint is in addition labeled as background dot, and other is labeled as pixel to be determined.
5. high-precision tongue body automatic division method according to claim 1, is characterized in that, the concrete steps of described step (4) are as follows:
(41) travel through all pixels to be identified, to each point, utilize pixel in the neighborhood of place 3*3 to set up linear color estimation model to it:
(411) the corresponding weight vector of pixel is solved:
In above formula, the pixel number of image is designated as
n, for label be
ipixel,
imeet
1<i<n, record its color value
; The color matrix recording all pixels in its neighborhood is
,
ifor unit matrix, coefficient
tvalue is 0.000001,
f i for required weight vector; Namely for pixel
i, its pixel color value can use weight vector
f i with its neighborhood territory pixel color value Linearly Representation;
(412) weight vector expansion: namely for pixel
i, its pixel color value can use weight vector
ξ i with pixel color value Linearly Representation in all images:
If wherein pixel
jfor pixel
ineighbor, its value is weight vector
f i middle respective value, otherwise be 0;
(42) image configuration global restriction equation is solved, obtain all foreground pixels point, complete segmentation;
(421) image configuration global restriction equation is solved, obtains foreground area masking-out, value be 1 point be foreground point, value be-1 point be background dot:
In above formula, weight matrix
, size is
n*n;
cfor diagonal matrix, if pixel
jfor marked pixels, then diagonal matrix
c?
jrow diagonal element value is constant 100000, otherwise value is 0;
α * for the classified information vector of marker image vegetarian refreshments, foreground point value 1, background dot value-1;
(422) will
αmiddle value is the pixel extraction of 1 and preserves into the new image of a width according to its former coordinate, completes tongue body segmentation.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410823935.XA CN104537379A (en) | 2014-12-26 | 2014-12-26 | High-precision automatic tongue partition method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410823935.XA CN104537379A (en) | 2014-12-26 | 2014-12-26 | High-precision automatic tongue partition method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN104537379A true CN104537379A (en) | 2015-04-22 |
Family
ID=52852899
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410823935.XA Pending CN104537379A (en) | 2014-12-26 | 2014-12-26 | High-precision automatic tongue partition method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104537379A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105930798A (en) * | 2016-04-21 | 2016-09-07 | 厦门快商通科技股份有限公司 | Tongue image quick detection and segmentation method based on learning and oriented to handset application |
CN106327506A (en) * | 2016-08-05 | 2017-01-11 | 北京三体高创科技有限公司 | Probability-partition-merging-based three-dimensional model segmentation method |
CN108615239A (en) * | 2018-05-10 | 2018-10-02 | 福建中医药大学 | Tongue image dividing method based on threshold technology and Gray Projection |
CN109325387A (en) * | 2017-07-31 | 2019-02-12 | 株式会社理光 | Image processing method, device, electronic equipment |
CN109598297A (en) * | 2018-11-29 | 2019-04-09 | 新绎健康科技有限公司 | A kind of tongue fur tongue analysis method, system, computer equipment and storage medium |
CN109636864A (en) * | 2018-12-19 | 2019-04-16 | 新绎健康科技有限公司 | A kind of tongue dividing method and system based on color correction Yu depth convolutional neural networks |
CN109740611A (en) * | 2019-01-25 | 2019-05-10 | 中电健康云科技有限公司 | Tongue image analysis method and device |
CN110599463A (en) * | 2019-08-26 | 2019-12-20 | 依脉人工智能医疗科技(天津)有限公司 | Tongue image detection and positioning algorithm based on lightweight cascade neural network |
WO2020108437A1 (en) * | 2018-11-26 | 2020-06-04 | 深圳市前海安测信息技术有限公司 | Sublingual vein feature extraction apparatus and method |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1367455A (en) * | 2002-03-25 | 2002-09-04 | 北京工业大学 | Chinese medicine tongue image computer cutting based on sample strip Snakes model |
CN101799870A (en) * | 2009-02-06 | 2010-08-11 | 天津市医学堂科技有限公司 | Tongue picture profile extracting method and application thereof |
CN103761506A (en) * | 2014-01-02 | 2014-04-30 | 上海大学 | Method for recognizing fissured tongue based on support vector machine |
-
2014
- 2014-12-26 CN CN201410823935.XA patent/CN104537379A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1367455A (en) * | 2002-03-25 | 2002-09-04 | 北京工业大学 | Chinese medicine tongue image computer cutting based on sample strip Snakes model |
CN101799870A (en) * | 2009-02-06 | 2010-08-11 | 天津市医学堂科技有限公司 | Tongue picture profile extracting method and application thereof |
CN103761506A (en) * | 2014-01-02 | 2014-04-30 | 上海大学 | Method for recognizing fissured tongue based on support vector machine |
Non-Patent Citations (3)
Title |
---|
XIAOQIANG LI ET AL.: ""Automatic Tongue Image Segmentation Based on Histogram projection and matting"", 《2014 IEEE INTERNATIONAL CONFERENCE ON BIOINFORMATICS AND BIOMEDICINE》 * |
张灵 等: ""基于灰度投影和阈值自动选取的舌像分割方法"", 《中国组织工程研究与临床康复》 * |
王艳清 等: ""一种基于先验知识的自动舌体分割算法"", 《电子学报》 * |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105930798B (en) * | 2016-04-21 | 2019-05-03 | 厦门快商通科技股份有限公司 | The tongue picture towards mobile phone application based on study quickly detects dividing method |
CN105930798A (en) * | 2016-04-21 | 2016-09-07 | 厦门快商通科技股份有限公司 | Tongue image quick detection and segmentation method based on learning and oriented to handset application |
CN106327506A (en) * | 2016-08-05 | 2017-01-11 | 北京三体高创科技有限公司 | Probability-partition-merging-based three-dimensional model segmentation method |
CN106327506B (en) * | 2016-08-05 | 2019-11-08 | 北京三体高创科技有限公司 | A kind of threedimensional model dividing method merged based on probability subregion |
CN109325387A (en) * | 2017-07-31 | 2019-02-12 | 株式会社理光 | Image processing method, device, electronic equipment |
CN109325387B (en) * | 2017-07-31 | 2021-09-28 | 株式会社理光 | Image processing method and device and electronic equipment |
CN108615239A (en) * | 2018-05-10 | 2018-10-02 | 福建中医药大学 | Tongue image dividing method based on threshold technology and Gray Projection |
CN108615239B (en) * | 2018-05-10 | 2020-03-27 | 福建中医药大学 | Tongue image segmentation method based on threshold technology and gray level projection |
WO2020108437A1 (en) * | 2018-11-26 | 2020-06-04 | 深圳市前海安测信息技术有限公司 | Sublingual vein feature extraction apparatus and method |
CN109598297A (en) * | 2018-11-29 | 2019-04-09 | 新绎健康科技有限公司 | A kind of tongue fur tongue analysis method, system, computer equipment and storage medium |
CN109636864A (en) * | 2018-12-19 | 2019-04-16 | 新绎健康科技有限公司 | A kind of tongue dividing method and system based on color correction Yu depth convolutional neural networks |
CN109740611A (en) * | 2019-01-25 | 2019-05-10 | 中电健康云科技有限公司 | Tongue image analysis method and device |
CN110599463A (en) * | 2019-08-26 | 2019-12-20 | 依脉人工智能医疗科技(天津)有限公司 | Tongue image detection and positioning algorithm based on lightweight cascade neural network |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104537379A (en) | High-precision automatic tongue partition method | |
CN113240691B (en) | Medical image segmentation method based on U-shaped network | |
CN107665492B (en) | Colorectal panoramic digital pathological image tissue segmentation method based on depth network | |
CN111145209B (en) | Medical image segmentation method, device, equipment and storage medium | |
CN107862698A (en) | Light field foreground segmentation method and device based on K mean cluster | |
CN107730507A (en) | A kind of lesion region automatic division method based on deep learning | |
CN107346550B (en) | It is a kind of for the three dimensional point cloud rapid registering method with colouring information | |
CN106503695B (en) | A kind of tobacco plant identification and method of counting based on Aerial Images | |
CN101710418B (en) | Interactive mode image partitioning method based on geodesic distance | |
CN112184705A (en) | Human body acupuncture point identification, positioning and application system based on computer vision technology | |
US20150186755A1 (en) | Systems and Methods for Object Identification | |
CN105574527A (en) | Quick object detection method based on local feature learning | |
CN106157279A (en) | Eye fundus image lesion detection method based on morphological segment | |
CN110458835A (en) | A kind of image processing method, device, equipment and medium | |
CN109253722A (en) | Merge monocular range-measurement system, method, equipment and the storage medium of semantic segmentation | |
CN104517101A (en) | Game poker card recognition method based on pixel square difference matching | |
CN104021566A (en) | GrabCut algorithm-based automatic segmentation method of tongue diagnosis image | |
CN109978895B (en) | Tongue image segmentation method and device | |
CN106447673A (en) | Chip pin extraction method under non-uniform illumination condition | |
CN113706562B (en) | Image segmentation method, device and system and cell segmentation method | |
Rachna et al. | Detection of Tuberculosis bacilli using image processing techniques | |
CN109325451A (en) | A kind of high spatial resolution arable land full-automatic extractive technique in plot based on deep learning | |
CN110400287B (en) | Colorectal cancer IHC staining image tumor invasion edge and center detection system and method | |
CN111326238A (en) | Cancer cell detection device based on sliding window | |
CN113237881B (en) | Detection method and device for specific cells and pathological section detection system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20150422 |
|
WD01 | Invention patent application deemed withdrawn after publication |