CN107292347A - A kind of capsule endoscope image-recognizing method - Google Patents
A kind of capsule endoscope image-recognizing method Download PDFInfo
- Publication number
- CN107292347A CN107292347A CN201710544884.0A CN201710544884A CN107292347A CN 107292347 A CN107292347 A CN 107292347A CN 201710544884 A CN201710544884 A CN 201710544884A CN 107292347 A CN107292347 A CN 107292347A
- Authority
- CN
- China
- Prior art keywords
- image
- grader
- capsule endoscope
- classification
- color
- 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
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/337—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving reference images or patches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Quality & Reliability (AREA)
- Artificial Intelligence (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Endoscopes (AREA)
Abstract
The invention discloses a kind of capsule endoscope image-recognizing method, comprise the following steps：A. image is obtained；B. image preprocessing；C. color of image feature extraction；D. image texture characteristic is extracted；E. multi-features；F. image classification is handled.The method of the invention is based on color characteristic and textural characteristics, image is identified and classified by way of carrying out machine learning based on the grader of algorithm of support vector machine thought, the degree of accuracy is high, classification is clear and definite, it may determine that complicated image, the photo of bleeding and lesion can effectively be identified, reduce the workload of medical worker, shorten the observing time of medical worker, the medical diagnosis reference of science is made for medical worker, avoid because photo is more, scoring time short-range missile causes to fail to pinpoint a disease in diagnosis even mistake and examines the generation of phenomenon, and then has influence on the treatment to patient.
Description
Technical field
The invention belongs to Computer Applied Technology field, and in particular to a kind of capsule endoscope image-recognizing method.
Background technology
Small intestine not still alimentary canal most long hollow organ (3.35-7.85m), is also maximum dirty of most bending, mobility
Device.At present except subregion can be with addition to conventional gastroscope colonoscopy, other parts are most difficult in whole gastrointestinal examination
The part of arrival, therefore the diagnosis and treatment of disease of intestine far lag behind other positions of intestines and stomach.The invention pole of capsule endoscope
Big this problem of having taken on a new look.Capsule endoscope is moved in alimentary canal, can obtain the figure of complete small intestinal segment in sufferer alimentary canal
Picture, breaches the blind area of small bowel examination, substantially increase disease of digestive tract diagnosis recall rate, auxiliary doctor make more science and
Effectively diagnosis and further treatment.
But capsule endoscope can only be moved with the wriggling of enteron aisle, and intestines peristalsis is general relatively slow, causes in capsule
Sight glass inspection is extracted the alimentary canal image of tens of thousands of, wherein actual lesion image is typically less than the 1% of its sum.And absolutely
Most image is all normal alimentary canal image, but doctor stills need to take a long time and browsed, and therefrom chooses
Bleeding and the image of lesion, carry out repeating diagosis, this is the work extremely wasted time and energy while also needing to two doctors.This
It is outer because doctor will be observed multiple images simultaneously, add observing time short, it is easy to cause to fail to pinpoint a disease in diagnosis even mistake examine from
And have influence on the treatment to patient.
The content of the invention
The technical problem to be solved in the present invention is：A kind of capsule endoscope image-recognizing method is provided, methods described is based on
Color characteristic and textural characteristics, carry out image recognition and calssification by way of grader carries out machine learning, and the degree of accuracy is high,
Classification is clear and definite, it can be determined that complicated image, can effectively identify the photo of bleeding and lesion, reduces the work of medical worker
Amount, shortens the observing time of medical worker, is the medical diagnosis reference that medical worker makes science, it is to avoid because photo is more, diagosis
Time short-range missile causes to fail to pinpoint a disease in diagnosis even mistake and examines the generation of phenomenon, and then has influence on the treatment to patient.
The technical scheme that the present invention solves technical problem is as follows：
The present invention is a kind of capsule endoscope image-recognizing method, is comprised the following steps：
A. image is obtained：Obtain the image that capsule endoscope is shot；
B. image preprocessing：The image got is pre-processed and effective photo is picked out, while deleting invalid
Photo；
C. color of image feature extraction：Use color Moment Methods under hsv color space extract the color characteristic of image；
D. image texture characteristic is extracted：The textural characteristics of image are extracted under hsv color space；
E. multi-features：The step c color characteristics extracted and step the d textural characteristics extracted are merged and obtained
Obtain new feature；
F. image classification is handled：Classification based training is carried out to grader using the new feature after fusion, the classification is then utilized
Device is classified to pretreated photo and identifies bleeding photo and the photo of lesion.
Further, image preprocessing includes image enhaucament and image scene classification in the step b.
Further, color of image feature extraction uses three rank colors of the image under hsv color space in the step c
Away from being used as color characteristic.
Further, it is using Contourlet changes under hsv color space that image texture characteristic, which is extracted, in the step d
Change the textural characteristics for extracting image.
Further, classification is carried out to pretreatment image in the step f to refer to image being divided into invalid image, normogram
Picture, bleeding image and lesion image.
Further, classification based training is carried out to grader using the new feature after fusion in the step f, the grader is adopted
Algorithm of support vector machine thought is used, it is comprised the following steps that：
1) endoscopic images are subjected to manual sort in advance, be then stored in different files；
2) mask images specified in endoscopic images are read；
3) the classified endoscopic images in each file are read, using the mask images in previous step to of all categories
Image is blocked and extracts the feature of all images, while being classified according to classification；
4) all original images are subjected to out of order processing, machine learning is carried out to grader with half characteristics of image, will
The characteristics of image input grader for being used to train carries out machine learning；Second half characteristics of image come verify grader effect and
Performance, will the grader that trains of characteristics of image input to be sorted classified and obtain result；
5) result of classification results and manual sort are contrasted, reaches the classification instruction that grader is completed after requirement
Practice.
Further, the step 1) in endoscopic images are subjected to manual sort in advance, it is invalid to refer to image being divided into
Image, normal picture, bleeding image and lesion image.
The beneficial effects of the invention are as follows：The present invention uses HSV color spaces color characteristic and texture characteristic extracting method pair
Capsule endoscope image is analyzed, and using the new feature after color characteristic and Texture Feature Fusion to based on SVMs
The grader of algorithm carries out machine learning to realize identification and classification of the grader to endoscopic images, it is to avoid manual sort knows
The problem of not wasting time and energy, helps doctor quickly to make the medical diagnosis of science；The present invention is using to color under HSV color spaces
Two kinds of features of feature and textural characteristics are identified and classified the degree of accuracy that single features identification can be avoided to occur not to image
It is high, classify it is indefinite, the problems such as complicated image can not be judged.
Brief description of the drawings
Fig. 1 is heretofore described capsule endoscope image-recognizing method flow chart；
Fig. 2 is heretofore described classifier training flow chart；
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation is described.
As shown in figure 1, a kind of capsule endoscope image-recognizing method of the invention, comprises the following steps：
A. image is obtained：Obtain the image that capsule endoscope is shot；The endoscopic images in people digest road are obtained first, are led to
Cross Wireless capsule endoscope to take pictures to the alimentary canal of patient, the photo photographed then is transmitted out into backstage is handled；
B. image preprocessing：The image got is pre-processed and effective photo is picked out, while deleting invalid
Photo；The present embodiment image preprocessing includes image enhaucament and scene classification, and the coloured image of capsule endoscope collection is being obtained
It will necessarily be produced during taking and degrade and degenerate, so needing to carry out enhancing processing to it, it is more suitable for doctor's observation and is examined
It is disconnected, while carrying out scene classification processing to image, help to filter out category in image by positioning the position of pylorus and ileocaecal sphineter
In the scene image of small intestine；
C. color of image feature extraction：Use color Moment Methods under hsv color space extract the color characteristic of image；
There are three channel components, respectively passage H (tone), passage S (saturation degree) and passage V (brightness) in hsv color space；In HSV
Under color space, the characteristic pattern of channel components is observed it will be seen that the characteristic value appearance of preceding 3 rank of H component color squares is small
The fluctuation of amplitude, but after the 3rd rank, the characteristic value of color moment and its exponent number are into positive correlation；In addition may be used also by characteristic pattern
To find out, the preceding 3 rank characteristic value of S components and V component color moment is changed greatly, and in 4~9 ranks due to regular presence, S points
Amount and the diversity factor of V component are little；Consider the changing factor of three channel components, the last selective extraction three of the present embodiment
Rank color away from characteristic value as color characteristic characteristic vector.
D. image texture characteristic is extracted：The textural characteristics of image are extracted under hsv color space；Normal picture and Abnormal Map
The textural characteristics of picture have part variation, it is possible to both textural characteristics are obtained by texture feature extraction step；
Contourlet transformation can realize the decomposition of any direction on any yardstick, be good at profile and directionality line in description image
Manage information；Endoscopic images are transformed under hsv color space by the present embodiment using contourlet transformation texture feature extraction,
Simultaneously for the texture information being described more fully on endoscopic images different directions, and in view of the dimension mistake of characteristic vector
Conference reduces classifying quality, and the step carries out the sub-band division in 8 directions to each Color Channel respectively, altogether 24 subbands；
3 rank squares are finally asked respectively to this all subband, obtained new feature vector (72 dimension) are regard as the texture feature vector of image；
E. multi-features：The step c color characteristics extracted and step the d textural characteristics extracted are merged and obtained
Obtain new feature；Color moment is merged with contourlet transformation, is the side that a kind of color characteristic is combined with textural characteristics
Method；The present embodiment carries out contourlet transformation under hsv color space to each Color Channel, then to the image after conversion
Three rank color moments are extracted respectively, finally obtain the new feature vector with color characteristic and textural characteristics, this feature vector is made
The foundation of machine learning and classification is carried out for grader.
F. image classification is handled：The grader based on algorithm of support vector machine thought is carried out using the new feature after fusion
Classification based training, is then classified to pretreated photo using the grader and identifies bleeding photo and the photograph of lesion
Piece.The grader is after classification based training, you can the image that the endoscope of preprocessed mistake is obtained is identified and classified, speed
Degree is fast, efficiency high.
Pretreatment image is carried out in the present embodiment, in the step f classification refer to image being divided into invalid image, it is normal
Image, bleeding image and lesion image.
Using the new feature after fusion to being divided based on algorithm of support vector machine thought in step f described in the present embodiment
Class device carries out classification based training, i.e., using linear classification method, as shown in Fig. 2 it is comprised the following steps that：
1) endoscopic images are subjected to manual sort in advance, be then stored in different files；Here to endoscope
The manual sort of image is that image is divided into invalid image, normal picture, bleeding image and lesion image；
2) mask images specified in endoscopic images are read；
3) the classified endoscopic images in each file are read, using the mask images in previous step to of all categories
Image is blocked and extracts the feature of all images, while being classified according to classification；
4) all original images are subjected to out of order processing, machine learning is carried out to grader with half characteristics of image, will
The characteristics of image input grader for being used to train carries out machine learning；Second half characteristics of image come verify grader effect and
Performance, will the grader that trains of characteristics of image input to be sorted classified and obtain result；
5) result of classification results and manual sort are contrasted, reaches the classification instruction that grader is completed after requirement
Practice.
In the present embodiment, the step 2) in mask (mask) be a kind of with selected image, figure or object, to place
The image (all or local) of reason is blocked, and is a kind of conventional one kind to control region or the processing procedure of image procossing
Image processing method；By reading the defined mask images, it is possible to reduce extract the complexity and workload during characteristics of image,
After the mask images are to needing image to be processed to block, we can just extract effective characteristics of image, exclude nothing
Imitate the interference of feature；The step 3) be by step 2) in regulation mask images the image in file is blocked simultaneously
Extract the effective characteristics of image of all endoscopic images, and machine learning as grader in subsequent step and classification according to
According to.
Although an embodiment of the present invention has been shown and described, for the ordinary skill in the art, can be with
A variety of changes, modification can be carried out to these embodiments, replace without departing from the principles and spirit of the present invention by understanding
And modification, the scope of the present invention is defined by the appended.
Claims (7)
1. a kind of capsule endoscope image-recognizing method, comprises the following steps：
A. image is obtained：Obtain the image that capsule endoscope is shot；
B. image preprocessing：The image got is pre-processed and effective photo is picked out, while deleting invalid photograph
Piece；
C. color of image feature extraction：Use color Moment Methods under hsv color space extract the color characteristic of image；
D. image texture characteristic is extracted：The textural characteristics of image are extracted under hsv color space；
E. multi-features：The step c color characteristics extracted and step the d textural characteristics extracted are merged and obtain new
Feature；
F. image classification is handled：Classification based training is carried out to grader using the new feature after fusion, the grader pair is then utilized
Pretreated photo is classified and identifies bleeding photo and the photo of lesion.
2. a kind of capsule endoscope image-recognizing method according to claim 1, it is characterised in that：Scheme in the step b
As pretreatment includes image enhaucament and image scene classification.
3. a kind of capsule endoscope image-recognizing method according to claim 1, it is characterised in that：Scheme in the step c
As color feature extracted using three rank colors of the image under hsv color space away from being used as color characteristic.
4. a kind of capsule endoscope image-recognizing method according to claim 1, it is characterised in that：Scheme in the step d
As texture feature extraction is to extract the textural characteristics of image using contourlet transformation under hsv color space.
5. a kind of capsule endoscope image-recognizing method according to claim 1, it is characterised in that：It is right in the step f
Pretreatment image carries out classification and refers to image being divided into invalid image, normal picture, bleeding image and lesion image.
6. a kind of capsule endoscope image-recognizing method according to claim 1, it is characterised in that：Adopted in the step f
Classification based training is carried out to grader with the new feature after fusion, the grader uses algorithm of support vector machine thought, its specific step
It is rapid as follows：
1) endoscopic images are subjected to manual sort in advance, be then stored in different files；
2) mask images specified in endoscopic images are read；
3) the classified endoscopic images in each file are read, using the mask images in previous step to image of all categories
The feature of all images is blocked and is extracted, while being classified according to classification；
4) all original images are subjected to out of order processing, machine learning is carried out to grader with half characteristics of image, will the use
Machine learning is carried out in the characteristics of image input grader of training；Second half characteristics of image verifies the effect and property of grader
Can, will the grader that trains of characteristics of image input to be sorted classified and obtain result；
5) result of classification results and manual sort are contrasted, reaches the classification based training that grader is completed after requirement.
7. a kind of capsule endoscope image-recognizing method according to claim 6, it is characterised in that：The step 1) middle general
Endoscopic images carry out manual sort in advance, refer to image being divided into invalid image, normal picture, bleeding image and lesion figure
Picture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710544884.0A CN107292347A (en) | 2017-07-06 | 2017-07-06 | A kind of capsule endoscope image-recognizing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710544884.0A CN107292347A (en) | 2017-07-06 | 2017-07-06 | A kind of capsule endoscope image-recognizing method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107292347A true CN107292347A (en) | 2017-10-24 |
Family
ID=60100757
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710544884.0A Pending CN107292347A (en) | 2017-07-06 | 2017-07-06 | A kind of capsule endoscope image-recognizing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107292347A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108055454A (en) * | 2017-12-08 | 2018-05-18 | 合肥工业大学 | The architectural framework and image processing method of medical endoscope artificial intelligence chip |
| CN109091098A (en) * | 2017-10-27 | 2018-12-28 | 重庆金山医疗器械有限公司 | Magnetic control capsule endoscopic diagnostic and examination system |
| CN109241963A (en) * | 2018-08-06 | 2019-01-18 | 浙江大学 | Blutpunkte intelligent identification Method in capsule gastroscope image based on Adaboost machine learning |
| CN109635871A (en) * | 2018-12-12 | 2019-04-16 | 浙江工业大学 | A kind of capsule endoscope image classification method based on multi-feature fusion |
| CN110020610A (en) * | 2019-03-16 | 2019-07-16 | 复旦大学 | Colonoscopy quality examination control system based on deep learning |
| CN110084280A (en) * | 2019-03-29 | 2019-08-02 | 广州思德医疗科技有限公司 | A kind of method and device of determining tag along sort |
| CN110414607A (en) * | 2019-07-31 | 2019-11-05 | 中山大学 | Classification method, device, equipment and the medium of capsule endoscope image |
| CN111341441A (en) * | 2020-03-02 | 2020-06-26 | 刘四花 | Gastrointestinal disease model construction method and diagnosis system |
| CN111462082A (en) * | 2020-03-31 | 2020-07-28 | 重庆金山医疗技术研究院有限公司 | Focus picture recognition device, method and equipment and readable storage medium |
| CN113920042A (en) * | 2021-09-24 | 2022-01-11 | 深圳市资福医疗技术有限公司 | Image processing system and capsule endoscope |
| CN117572628A (en) * | 2024-01-02 | 2024-02-20 | 首都医科大学附属北京天坛医院 | Portable endoscope and application method thereof |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105512612A (en) * | 2015-11-24 | 2016-04-20 | 中科院合肥技术创新工程院 | SVM-based image classification method for capsule endoscope |
| CN106097335A (en) * | 2016-06-08 | 2016-11-09 | 安翰光电技术（武汉）有限公司 | Digestive tract focus image identification system and recognition methods |
-
2017
- 2017-07-06 CN CN201710544884.0A patent/CN107292347A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105512612A (en) * | 2015-11-24 | 2016-04-20 | 中科院合肥技术创新工程院 | SVM-based image classification method for capsule endoscope |
| CN106097335A (en) * | 2016-06-08 | 2016-11-09 | 安翰光电技术（武汉）有限公司 | Digestive tract focus image identification system and recognition methods |
Non-Patent Citations (1)
| Title |
|---|
| 宋余庆等: "基于多特征融合的医学图像识别研究", 《计算机应用研究》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109091098A (en) * | 2017-10-27 | 2018-12-28 | 重庆金山医疗器械有限公司 | Magnetic control capsule endoscopic diagnostic and examination system |
| CN108055454A (en) * | 2017-12-08 | 2018-05-18 | 合肥工业大学 | The architectural framework and image processing method of medical endoscope artificial intelligence chip |
| CN108055454B (en) * | 2017-12-08 | 2020-07-28 | 合肥工业大学 | System architecture of medical endoscope artificial intelligence chip and image processing method |
| CN109241963A (en) * | 2018-08-06 | 2019-01-18 | 浙江大学 | Blutpunkte intelligent identification Method in capsule gastroscope image based on Adaboost machine learning |
| CN109635871A (en) * | 2018-12-12 | 2019-04-16 | 浙江工业大学 | A kind of capsule endoscope image classification method based on multi-feature fusion |
| CN109635871B (en) * | 2018-12-12 | 2021-06-18 | 浙江工业大学 | Capsule endoscope image classification method based on multi-feature fusion |
| CN110020610B (en) * | 2019-03-16 | 2023-02-10 | 复旦大学 | Enteroscope quality inspection control system based on deep learning |
| CN110020610A (en) * | 2019-03-16 | 2019-07-16 | 复旦大学 | Colonoscopy quality examination control system based on deep learning |
| CN110084280A (en) * | 2019-03-29 | 2019-08-02 | 广州思德医疗科技有限公司 | A kind of method and device of determining tag along sort |
| CN110084280B (en) * | 2019-03-29 | 2021-08-31 | 广州思德医疗科技有限公司 | Method and device for determining classification label |
| CN110414607A (en) * | 2019-07-31 | 2019-11-05 | 中山大学 | Classification method, device, equipment and the medium of capsule endoscope image |
| CN111341441A (en) * | 2020-03-02 | 2020-06-26 | 刘四花 | Gastrointestinal disease model construction method and diagnosis system |
| CN111462082A (en) * | 2020-03-31 | 2020-07-28 | 重庆金山医疗技术研究院有限公司 | Focus picture recognition device, method and equipment and readable storage medium |
| CN113920042A (en) * | 2021-09-24 | 2022-01-11 | 深圳市资福医疗技术有限公司 | Image processing system and capsule endoscope |
| CN113920042B (en) * | 2021-09-24 | 2023-04-18 | 深圳市资福医疗技术有限公司 | Image processing system and capsule endoscope |
| CN117572628A (en) * | 2024-01-02 | 2024-02-20 | 首都医科大学附属北京天坛医院 | Portable endoscope and application method thereof |
| CN117572628B (en) * | 2024-01-02 | 2024-05-07 | 首都医科大学附属北京天坛医院 | Portable endoscope and application method thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107292347A (en) | A kind of capsule endoscope image-recognizing method | |
| JP7335552B2 (en) | Diagnostic imaging support device, learned model, operating method of diagnostic imaging support device, and diagnostic imaging support program | |
| CN106934799B (en) | Capsule endoscope visual aids diagosis system and method | |
| CN110600122B (en) | Digestive tract image processing method and device and medical system | |
| WO2020071677A1 (en) | Method and apparatus for diagnosing gastric lesions by using deep learning on gastroscopy images | |
| CN107730489A (en) | Wireless capsule endoscope small intestine disease variant computer assisted detection system and detection method | |
| CN110189303B (en) | NBI image processing method based on deep learning and image enhancement and application thereof | |
| CN109118485A (en) | Digestive endoscope image classification based on multitask neural network cancer detection system early | |
| CN107767365A (en) | A kind of endoscopic images processing method and system | |
| CN110727097A (en) | Pathological microscopic image real-time acquisition and analysis system, method, device and medium | |
| CN105657580B (en) | A kind of capsule endoscope video abstraction generating method | |
| KR20200101540A (en) | Smart skin disease discrimination platform system constituting API engine for discrimination of skin disease using artificial intelligence deep run based on skin image | |
| CN110852396A (en) | Sample data processing method for cervical image | |
| CN109934276A (en) | Capsule endoscope image classification system and method based on transfer learning | |
| CN109635871A (en) | A kind of capsule endoscope image classification method based on multi-feature fusion | |
| CN104203075B (en) | Medical image processing device | |
| CN104869884A (en) | Medical image processing device and medical image processing method | |
| CN113129287A (en) | Automatic lesion mapping method for upper gastrointestinal endoscope image | |
| CN109241963B (en) | Adaboost machine learning-based intelligent identification method for bleeding point in capsule gastroscope image | |
| CN115018767A (en) | Cross-modal endoscope image conversion and lesion segmentation method based on eigen expression learning | |
| CN118351123B (en) | Multi-feature fusion endoscope image quality evaluation method | |
| CN115049666A (en) | Endoscope virtual biopsy device based on color wavelet covariance depth map model | |
| CN114782760A (en) | Stomach disease picture classification system based on multitask learning | |
| CN111784669B (en) | Multi-range detection method for capsule endoscopic images | |
| CN111476312B (en) | Method for classifying lesion images based on convolutional neural network |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171024 |