CN109389610A - A kind of skin lesion area computation method based on artificial intelligence identification - Google Patents
A kind of skin lesion area computation method based on artificial intelligence identification Download PDFInfo
- Publication number
- CN109389610A CN109389610A CN201811388435.2A CN201811388435A CN109389610A CN 109389610 A CN109389610 A CN 109389610A CN 201811388435 A CN201811388435 A CN 201811388435A CN 109389610 A CN109389610 A CN 109389610A
- Authority
- CN
- China
- Prior art keywords
- lesion
- region
- picture
- lesion region
- artificial intelligence
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- 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
- G06T2207/30088—Skin; Dermal
Abstract
A kind of skin lesion area computation method based on artificial intelligence identification, the areal analysis method including recognition methods and lesion region to lesion region, the recognition methods of the lesion region include to identify lesion region from lesion picture;By the acquisition of data to skin lesion image, intelligence computation goes out skin lesion area, and it is accurate to calculate, and cooperation APP or PC is in use, simple and convenient, convenient for promoting.
Description
Technical field
The present invention relates to skin lesion areal calculation field, especially a kind of skin lesion areal calculation based on artificial intelligence identification
Method.
Background technique
The size of skin lesion area has great importance for clinical treatment decision in many cases, such as burn surface
Product size decides assessment and amount infused to burn grade;The face of the skin lesions such as psoriasis, leucoderma, atopic dermatitis
Product decides the usage amount of externally applied drug.However, at present mostly using skin lesion area with sufferers themselves' palm (including palmar surfaces of fingers)
Its area is the 1% of the body surface gross area, calculates small area skin lesion with this;Or according to the total body surface of body various pieces substantially Zhan
The ratio of area calculates larger skin lesion area, such as in the 100% body surface gross area: incidence accounts for 9% (head, face
3%) portion, neck respectively account for;Double upper limbs account for 18% (double upper arm 7%, double forearms 6%, both hands 5%);It is accounted for before and after trunk including perineum
27% (preceding body 13%, rear quarters 13%, perineum 1%);Double lower limb (containing buttocks) accounts for 46% (double sterns 5%, both thighs 21%, double small
Leg 13%, biped 7%) (women biped and stern respectively account for 6%).Both methods is difficult to accurately calculate skin lesion area, and
It is very cumbersome, it is not easy to remember, thus need to design a kind of accurate calculation method of calculating skin lesion area.
The present invention be exactly in order to solve problem above and carry out improvement.
Summary of the invention
The object of the present invention is to provide a kind of data acquisition by skin lesion image, intelligence computation goes out skin lesion area, letter
Folk prescription just, the skin lesion area computation method convenient for popularization based on artificial intelligence identification.
The present invention is that technical solution used by solving its technical problem is:
A kind of skin lesion area computation method based on artificial intelligence identification, including the recognition methods and disease to lesion region
Become the areal analysis method in region, the recognition methods of the lesion region includes to identify lesion region from lesion picture;
The identification step of the lesion region is as follows:
S1, initialization segmentation is carried out to lesion picture;
S2, the parameter or object of reference of lesion picture are analyzed;
S3, lesion region is determined;
The areal analysis method of the lesion region includes the areal analysis to lesion region, and its step are as follows:
Y1, analysis lesion region, generate lesion region profile scatterplot;
Y2, it is compared and analyzed with object of reference;
Y3, lesion region area is calculated;
Further, scale markers of the object of reference in the step S2 as lesion region in step S3;
Further, scale markers are not present in the lesion region;
The described lesion region can be identified from lesion picture under artificial guidance;
The described lesion region can not be identified from lesion picture under artificial guidance;
Specifically, the lesion picture is shot by depth camera, depth camera can obtain the parameter of lesion picture
And depth map figure is generated, the object of reference is depth map figure;
The lesion picture is shot by monocular cam, and the paster of a known dimensions, the reference are used when shooting
Object is the paster.
Wherein, the depth detection of depth camera is to calculate basic and focus a task in machine vision, quasi-
Target really is detected, may also need to do many image segmentations, is identified, the work of aspect is tracked, itself does not have depth detection function
The camera of energy, the principle (stereo, MVS) that stereoscopic vision can be used carrys out estimating depth, and has the camera shooting of depth detection
Head, such as Kinect also often seek depth using principle of parallax, project a pattern, then compare.
System subscriber terminal is mobile phone or PC, carries out permission control and flow control by Gate way cluster, goes forward side by side
Row delivery of services, the analysis carried out to picture are forwarded to picture analyzing cluster, and Machine learning cluster carries out mould daily
The upgrading of type iteration.
Working principle are as follows: for monocular cam, when shooting need to use the paster object of reference (reference of a known dimensions
Object is preferably able to identical with the width of lesion or height) it is placed in photo, photograph is then being uploaded by cell phone application or the end PC
Piece;System has an initial segmentation model, and after upload pictures, system can be generated preliminary user with this initial model
Lesion region, simultaneity factor can provide region modification function, and user such as has objection to the lesion region of identification, can be direct
Modifier area carries out final areal calculation according to object of reference after user confirms region and obtains lesion area;
For depth camera, can not have to place object of reference in imaging area, parameter when record is taken pictures when taking pictures,
Photo is analyzed to obtain lesion region, using the parameters such as the focal length of photo and the time difference can with the actual size of calculating foci, from
And carry out lesion area calculating.
The beneficial effects of the present invention are: by the acquisition of data to skin lesion image and to the comparative analysis of object of reference, point
Reference record when cutting or taking pictures, can be by the real area for accurately calculating intelligence and obtaining skin lesion to curved surface, and it is accurate to calculate,
It is brought conveniently to the treatment of clinically fire victim, APP or PC is in use, simple and convenient for cooperation, convenient for promoting.
Specific embodiment
In order to be easy to understand the technical means, the creative features, the aims and the efficiencies achieved by the present invention, below
In conjunction with specific embodiments, the present invention is further explained.
This it is a kind of based on artificial intelligence identification skin lesion area computation method, including to lesion region recognition methods and
The areal analysis method of lesion region, the recognition methods of the lesion region include to identify lesion region from lesion picture;
The identification step of the lesion region is as follows:
S1, initialization segmentation is carried out to lesion picture;
S2, the parameter or object of reference of lesion picture are analyzed;
S3, lesion region is determined;
The areal analysis method of the lesion region includes the areal analysis to lesion region, and its step are as follows:
Y1, analysis lesion region, are generated lesion region profile scatterplot, are analyzed using pre-training deep learning parted pattern
Uploading pictures lesion region generates lesion region profile scatterplot;
Acquisition common photo is shot using monocular cam to provide using user terminal first in the analysis of common photo
Image rotating function rotate picture so that picture level, every a line image data curvature is substantially similar from top to bottom;
Y2, it is compared and analyzed with object of reference, then uses zoom function, the image analysis frame that cooperation user terminal provides
In scale function, adjust position of the object of reference relative to scale, make to meet actual size, according to image resolution ratio and scale,
The corresponding physical length of image each pixel can be calculated;
Y3, lesion region area is calculated, the corresponding physical length of image each pixel that step Y2 kind is calculated
Mm/px is denoted as μ;
Then image area is analyzed, image level direction is divided into n equal part, each equal part can be seen in n infinity
A rectangle is done, the flat object of reference side length of raw water is a, and the width of the n-th equal part is Δ d, a length of a ' of the n-th equal part, width in image
For Δ d ', then former object of reference area Sn=a × Δ d, object of reference area S'n=a' × Δ d' on image, then lesion region
Area be exactlyShooting angle be basically parallel to lesion cut be closer when because perspective generate
Distort very little, it is believed that under such condition, a ' and a are essentially identical, and then lesion region real area is exactlyS ' is the elemental area of object of reference;
In another case, lesion region area is larger, has been more than the lower regions of object of reference calibration, can be used
The data of object of reference part are fitted one, in vertical direction cube polynomial curve: ay3+by2+ cy+d=z, then basis
Area of the curved surface on shooting photo, calculates Δ d and Δ d ', finally calculates real area, finally calculate area
Further, scale markers of the object of reference in the step S2 as lesion region in step S3;
Further, scale markers are not present in the lesion region;
The described lesion region can be identified from lesion picture under artificial guidance;
The described lesion region can not be identified from lesion picture under artificial guidance;
Specifically, the lesion picture is shot by depth camera, depth camera can obtain the parameter of lesion picture
And depth map figure is generated, the object of reference is depth map figure, when using depth camera, can not had in imaging area
Middle placement object of reference, parameter when record is taken pictures when taking pictures, can obtain depth map figure according to the characteristic of depth camera,
Photo is analyzed to obtain lesion region, using the parameters such as the focal length of photo and the time difference can with the actual size of calculating foci, from
And lesion area can be more accurately obtained, lesion identification, calculating foci region area Spx are carried out to common photo2, so
It is aligned and is registrated with depth map figure afterwards, calculated the area that every square indicates, finally calculate area;
The lesion picture is shot by monocular cam, and the paster of a known dimensions, the reference are used when shooting
Object be the paster, it is known that the paster object of reference of size be preferably able to the width of lesion or height it is identical be placed in photo, then
Passing through cell phone application or the end PC upload pictures;System has an initial segmentation model, and user is after upload pictures, system
Preliminary lesion region can be generated with this initial model, simultaneity factor can provide region modification function, and user is such as to identification
Lesion region have objection, can direct modifier area, user confirm region after final face is carried out according to object of reference
Lesion area is calculated in product.
Wherein, the depth detection of depth camera is to calculate basic and focus a task in machine vision, quasi-
Target really is detected, may also need to do many image segmentations, is identified, the work of aspect is tracked, itself does not have depth detection function
The camera of energy, the principle (stereo, MVS) that stereoscopic vision can be used carrys out estimating depth, and has the camera shooting of depth detection
Head, such as Kinect also often seek depth using principle of parallax, project a pattern, then compare.
System subscriber terminal is mobile phone or PC, carries out permission control and flow control by Gate way cluster, goes forward side by side
Row delivery of services, the analysis carried out to picture are forwarded to picture analyzing cluster, and Machine learning cluster carries out mould daily
The upgrading of type iteration.
By the acquisition of data to skin lesion image and to the comparative analysis of object of reference, divide or take pictures when reference record,
Can be by the real area for accurately calculating intelligence and obtaining skin lesion to curved surface, it is accurate to calculate, to the treatment of clinically fire victim
It brings conveniently, APP or PC is in use, simple and convenient for cooperation, convenient for promoting.It has been shown and described above of the invention basic
Principle, main feature and advantages of the present invention.It should be understood by those skilled in the art that the present invention is not by above-described embodiment
Limitation, the above embodiments and description only illustrate the principle of the present invention, is not departing from spirit and scope of the invention
Under the premise of various changes and improvements may be made to the invention, these changes and improvements both fall within scope of the claimed invention
It is interior.The claimed scope of the invention is defined by appended claims and its equivalent.
Claims (7)
1. it is a kind of based on artificial intelligence identification skin lesion area computation method, including to lesion region recognition methods and diseased region
The areal analysis method in domain, it is characterised in that:
The recognition methods of the lesion region includes to identify lesion region from lesion picture;
The identification step of the lesion region is as follows:
S1, initialization segmentation is carried out to lesion picture;
S2, the parameter or object of reference of lesion picture are analyzed;
S3, lesion region is determined;
The areal analysis method of the lesion region includes the areal analysis to lesion region, and its step are as follows:
Y1, analysis lesion region, generate lesion region profile scatterplot;
Y2, it is compared and analyzed with object of reference;
Y3, lesion region area is calculated.
2. a kind of skin lesion area computation method based on artificial intelligence identification as described in claim 1, which is characterized in that described
Scale markers of the object of reference as lesion region in step S3 in step S2.
3. a kind of skin lesion area computation method based on artificial intelligence identification as described in claim 1, which is characterized in that described
Scale markers are not present in lesion region.
4. a kind of skin lesion area computation method based on artificial intelligence identification as claimed in claim 2, which is characterized in that described
The lesion region can be identified from lesion picture under artificial guidance.
5. a kind of skin lesion area computation method based on artificial intelligence identification as claimed in claim 2, which is characterized in that described
The lesion region can not be identified from lesion picture under artificial guidance.
6. a kind of skin lesion area computation method based on artificial intelligence identification as described in claim 1, which is characterized in that described
Lesion picture is shot by depth camera, and depth camera can obtain the parameter of lesion picture and generate depth map figure, institute
Stating object of reference is depth map figure.
7. a kind of skin lesion area computation method based on artificial intelligence identification as described in claim 1, which is characterized in that described
Lesion picture is shot by monocular cam, and using the paster of a known dimensions when shooting, the object of reference is the paster.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811388435.2A CN109389610A (en) | 2018-11-21 | 2018-11-21 | A kind of skin lesion area computation method based on artificial intelligence identification |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811388435.2A CN109389610A (en) | 2018-11-21 | 2018-11-21 | A kind of skin lesion area computation method based on artificial intelligence identification |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109389610A true CN109389610A (en) | 2019-02-26 |
Family
ID=65429658
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811388435.2A Pending CN109389610A (en) | 2018-11-21 | 2018-11-21 | A kind of skin lesion area computation method based on artificial intelligence identification |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109389610A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109961448A (en) * | 2019-04-10 | 2019-07-02 | 杭州智团信息技术有限公司 | Sketch the contours method and system in lesion tissue region |
CN110009630A (en) * | 2019-04-15 | 2019-07-12 | 中国医学科学院皮肤病医院 | A kind of skin targets region automatic testing method based on deep learning |
CN110335304A (en) * | 2019-06-11 | 2019-10-15 | 苏州思白人工智能技术研发有限公司 | Skin lesion area measurement method and skin disease diagnosis and therapy system based on image recognition |
CN112288686A (en) * | 2020-07-29 | 2021-01-29 | 深圳市智影医疗科技有限公司 | Model training method and device, electronic equipment and storage medium |
CN113257391A (en) * | 2021-06-02 | 2021-08-13 | 杭州咏柳科技有限公司 | Course of disease management system of skin disease |
CN114882098A (en) * | 2021-09-26 | 2022-08-09 | 上海交通大学医学院附属第九人民医院 | Method, system and readable storage medium for measuring area of specific region of living body |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070276309A1 (en) * | 2006-05-12 | 2007-11-29 | Kci Licensing, Inc. | Systems and methods for wound area management |
CN101966083A (en) * | 2010-04-08 | 2011-02-09 | 太阳系美容事业有限公司 | Abnormal skin area computing system and computing method |
US20110210961A1 (en) * | 2010-02-26 | 2011-09-01 | Clark Alexander Bendall | Method of determining the profile of a surface of an object |
US20140088402A1 (en) * | 2012-09-25 | 2014-03-27 | Innovative Therapies, Inc. | Wound measurement on smart phones |
CN107424157A (en) * | 2017-08-11 | 2017-12-01 | 南京航空航天大学 | Animal sticks the computational methods and computing system of contact zone real contact area |
-
2018
- 2018-11-21 CN CN201811388435.2A patent/CN109389610A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070276309A1 (en) * | 2006-05-12 | 2007-11-29 | Kci Licensing, Inc. | Systems and methods for wound area management |
CN101442970A (en) * | 2006-05-12 | 2009-05-27 | 凯希特许有限公司 | Systems and methods for wound area management |
US20110210961A1 (en) * | 2010-02-26 | 2011-09-01 | Clark Alexander Bendall | Method of determining the profile of a surface of an object |
CN101966083A (en) * | 2010-04-08 | 2011-02-09 | 太阳系美容事业有限公司 | Abnormal skin area computing system and computing method |
US20140088402A1 (en) * | 2012-09-25 | 2014-03-27 | Innovative Therapies, Inc. | Wound measurement on smart phones |
CN107424157A (en) * | 2017-08-11 | 2017-12-01 | 南京航空航天大学 | Animal sticks the computational methods and computing system of contact zone real contact area |
Non-Patent Citations (2)
Title |
---|
赵飞 等: "适用于创伤修复研究的伤口面积测量系统设计", 《医疗卫生装备》 * |
量子位: "语义分割中的深度学习方法全解:从FCN、SegNet 到各版本DeepLab", 《搜狐网》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109961448A (en) * | 2019-04-10 | 2019-07-02 | 杭州智团信息技术有限公司 | Sketch the contours method and system in lesion tissue region |
CN110009630A (en) * | 2019-04-15 | 2019-07-12 | 中国医学科学院皮肤病医院 | A kind of skin targets region automatic testing method based on deep learning |
CN110335304A (en) * | 2019-06-11 | 2019-10-15 | 苏州思白人工智能技术研发有限公司 | Skin lesion area measurement method and skin disease diagnosis and therapy system based on image recognition |
CN112288686A (en) * | 2020-07-29 | 2021-01-29 | 深圳市智影医疗科技有限公司 | Model training method and device, electronic equipment and storage medium |
CN112288686B (en) * | 2020-07-29 | 2023-12-19 | 深圳市智影医疗科技有限公司 | Model training method and device, electronic equipment and storage medium |
CN113257391A (en) * | 2021-06-02 | 2021-08-13 | 杭州咏柳科技有限公司 | Course of disease management system of skin disease |
CN114882098A (en) * | 2021-09-26 | 2022-08-09 | 上海交通大学医学院附属第九人民医院 | Method, system and readable storage medium for measuring area of specific region of living body |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109389610A (en) | A kind of skin lesion area computation method based on artificial intelligence identification | |
WO2021077720A1 (en) | Method, apparatus, and system for acquiring three-dimensional model of object, and electronic device | |
CN105894574B (en) | A kind of binocular three-dimensional reconstruction method | |
CN107403168B (en) | Face recognition system | |
CN106327571B (en) | A kind of three-dimensional face modeling method and device | |
CN104317391B (en) | A kind of three-dimensional palm gesture recognition exchange method and system based on stereoscopic vision | |
CN106251399A (en) | A kind of outdoor scene three-dimensional rebuilding method based on lsd slam | |
CN103971408B (en) | Three-dimensional facial model generating system and method | |
CN104680582A (en) | Method for creating object-oriented customized three-dimensional human body model | |
CN103366157B (en) | A kind of determination methods of human eye sight distance | |
US20190392564A1 (en) | Electronic Device, and Control Method and Control Apparatus for the Same | |
CN100573581C (en) | Semi-automatic partition method of lung CT image focus | |
CN101815174B (en) | Control method and control device for camera shooting | |
CN107831900B (en) | human-computer interaction method and system of eye-controlled mouse | |
CN103049758B (en) | Merge the remote auth method of gait light stream figure and head shoulder mean shape | |
CN106127696B (en) | A kind of image removal method for reflection based on BP neural network fitting sports ground | |
CN104573634A (en) | Three-dimensional face recognition method | |
CN113177977B (en) | Non-contact three-dimensional human body size measuring method | |
CN107944435A (en) | A kind of three-dimensional face identification method, device and processing terminal | |
CN103870808A (en) | Finger vein identification method | |
CN106650701A (en) | Binocular vision-based method and apparatus for detecting barrier in indoor shadow environment | |
CN113065502A (en) | 3D information acquisition system based on standardized setting | |
CN105761243A (en) | Three-dimensional full face photographing system based on structured light projection and photographing method thereof | |
CN109308462B (en) | Finger vein and knuckle print region-of-interest positioning method | |
CN104036541A (en) | Fast three-dimensional reconstruction method in vision measurement |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190226 |
|
WD01 | Invention patent application deemed withdrawn after publication |