CN108960042A - The echinococcus protoscolex survival rate detection method of vision significance and SIFT feature - Google Patents
The echinococcus protoscolex survival rate detection method of vision significance and SIFT feature Download PDFInfo
- Publication number
- CN108960042A CN108960042A CN201810476337.8A CN201810476337A CN108960042A CN 108960042 A CN108960042 A CN 108960042A CN 201810476337 A CN201810476337 A CN 201810476337A CN 108960042 A CN108960042 A CN 108960042A
- Authority
- CN
- China
- Prior art keywords
- protoscolex
- image
- sift feature
- echinococcus
- processing
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/69—Microscopic objects, e.g. biological cells or cellular parts
- G06V20/693—Acquisition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- 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
- 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
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/69—Microscopic objects, e.g. biological cells or cellular parts
- G06V20/695—Preprocessing, e.g. image segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/69—Microscopic objects, e.g. biological cells or cellular parts
- G06V20/698—Matching; Classification
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Probability & Statistics with Applications (AREA)
- Image Analysis (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The present invention relates to detection method technical fields, are the echinococcus protoscolex survival rate detection methods of a kind of vision significance and SIFT feature;Recognition methods the present invention is based on vision significance and SIFT feature is calculated image local, only consider the doubtful protoscolex region obtained by notable figure, the plenty of time is saved from SIFT feature extraction, in the acquisition of the SIFT feature vector of sample image, the present invention carries out cluster to it using k-means algorithm and changes SIFT feature vector there are the descriptions of the feature of higher-dimension, it is excessively complicated to solve calculating, the problems such as taking long time, improve the computational efficiency that target search is carried out using SIFT feature, SIFT feature in marking area is more stable, change the human error as caused by artificial counting, the artificial counting time is longer, caused by working efficiency it is low, the problem of heavy workload, the accuracy of identification is also ensured simultaneously.
Description
Technical field
The present invention relates to detection method technical fields, are the echinococcus protoscolexs of a kind of vision significance and SIFT feature
Survival rate detection method.
Background technique
Echinococcosis is the serious parasitic disease as caused by the larva of Echinococcus granulosus and Echinococcus multilocularis, to people and
Animal all has an impact, and this disease is most commonly seen in the generally existing area of animal husbandry.China belongs to the high-incidence country of echinococcosis
One, including Xinjiang Uygur Autonomous Regions, some areas are still in high fashion trend.At present in the prevention and treatment process of echinococcosis
In, treating echinococcosis treats pharmacodynamic assessment urgently to be speculated, makes evaluating drug effect and rationally utilizes needs further perfect, wherein body
It is outer to determine whether Echinococcus hydatid cyst is dead to the evaluating drug effect for researching and developing novel hydatid Drugs and treating echinococcosis and rationally using most important.
It is the external common method for determining the death of Echinococcus Granulosus Cysts protoscolex that dye method is refused in Yihong, has the advantages such as simple and economical, quick,
By in-vitro screening drug, researches and develops novel liver hydatid Drugs field and commonly use.It is eosin stains that the dyeing liquor that dye method uses is refused in Yihong
Liquid.Dye principle is refused using dyestuff, i.e., the related cell membrane based on damage, the film on non-live (dead) cell allows non-transmission
Film property dyestuff can enter dyeing in film;And the cell membrane of living cells can resist dyestuff entrance, dye phenomenon is refused in generation, not colored.
It refuses in strict accordance with protoscolex to survive in dead counting criteria after dye method export video and picture by Yihong under the microscope to calculate
Survival condition.Although Yihong refuse dye method simply and quickly, the advantages such as economy, have some disadvantages.Such as: artificial counting exists
Human error, artificial counting time are longer.Therefore it is necessary to identify that dye method image is refused in Yihong for developing intellectual resource.
Although current Yihong is refused dye method digital image recognition algorithm and had not been reported, conventional polypide identification field has
Algorithm, for example, Rema M. et al. is developed based on a kind of parasite of human ovum micro-image segmentation by active contour model
Number, effectively realizes the segmentation to worm's ovum image under complex background, and Chen Bili etc. proposes a kind of based on morphologic filtering method
Parasite egg pictures mixing partitioning algorithm, pass through the improvement to morphologic filtering and combine convex closure operation, to parasitic ovum
Image has carried out effective segmentation.High source etc. has used support vector machines as classifier, to the japonice ovum in micro-image
Identification is studied, and Zhang Zhenglong proposes that improved k nearest neighbor classifier carries out identification classification to it.In in recent years, for
Parasites identification technology in micro- medicine, many scholars at home and abroad have done a large amount of research, but existing method has identification
The part required manual intervention in treatment process is too many, and the characteristics of image of proposition cannot reflect feature of image, makes various identifications pair
The range of characteristic values overlapping of elephant is more etc., leads to low efficiency, heavy workload and there are problems that human error.
Summary of the invention
The present invention provides the echinococcus protoscolex survival rate detection method of a kind of vision significance and SIFT feature, gram
The deficiency for having taken the above-mentioned prior art, can effectively solve existing method, there are the portions required manual intervention in identification processing procedure
Dividing too much, the characteristics of image of proposition cannot reflect feature of image, keep the range of characteristic values overlapping of various identification objects more etc., it leads
It causes low efficiency, heavy workload and there are problems that human error.
The technical scheme is that realized by following measures: a kind of spine ball of vision significance and SIFT feature
Tapeworm protoscolex survival rate detection method carries out in the steps below: the first step, by 20 to be detected to 100 echinococcus originals
The head larva of a tapeworm or the cercaria of a schistosome refuses the processing of dye method by Yihong or Trypan Blue is handled, and after processing and takes pictures, obtains echinococcus protoscolex to be detected
Handle image;Second step, to the polypide image saliency map of echinococcus protoscolex to be detected processing image zooming-out color and brightness;
Third step carries out linear weighted function to the polypide image saliency map of color and brightness and generates total notable figure;4th step is extracted total significant
The marking area of figure finds the central point of doubtful polypide in marking area and cuts all doubtful living worm body slices, significant
These suspected target zone markers are come out in region, then by SIFT algorithm, extract the sift feature of living worm body slice, it is raw
At the sift feature vector of corresponding suspicious region;5th step, by the sift feature vector of suspicious region and echinococcus procephalon
In larva of a tapeworm or the cercaria of a schistosome living worm body image data figure cluster after sift feature vector identification is compared, if matching identification the result is that work worm
Body then cancels label if the result of matching identification is not living worm body labeled as living worm body target, will finally mark result
It is restored on echinococcus protoscolex processing image to be detected and counts, obtain echinococcus protoscolex survival rate.
Here is the further optimization and/or improvements to invention technology described above scheme:
Above-mentioned echinococcus protoscolex living worm body image data figure obtains in the steps below: the first step takes 20 to 100 spine balls
Tapeworm protoscolex refuses the processing of dye method by Yihong or Trypan Blue is handled, and after processing and takes pictures, obtains echinococcus protoscolex
Handle image;Second step, choose 50 width to 70 width echinococcus protoscolexs processing image in living worm body image and to worm of taking on service jobs
The background image of body image establishes database;Third step extracts living worm body image and Background in database by SIFT algorithm
The sift feature vector of picture;4th step clusters sift feature vector by k-means algorithm, then will be after cluster
Sift feature vector is put into svm classifier, obtains echinococcus protoscolex living worm body image data figure.
The echinococcus protoscolex survival rate detection method of vision significance of the present invention and SIFT feature is to original to be detected
Head larva of a tapeworm or the cercaria of a schistosome image is identified that protoscolex living substantially falls in labeled image-region;View-based access control model conspicuousness and SIFT are special
The recognition methods of sign is calculated image local, the doubtful protoscolex region obtained by notable figure is only considered, from SIFT
Feature extraction saves the plenty of time, and in the acquisition of the SIFT feature vector of sample image, the present invention uses k-means algorithm
Cluster is carried out to it and changes SIFT feature vector there are the description of the feature of higher-dimension, is solved and is calculated excessively complexity, takes long time
The problems such as, the computational efficiency that target search is carried out using SIFT feature is improved, the SIFT feature in marking area is more
Stablize, change the human error as caused by artificial counting, the artificial counting time is longer, caused by working efficiency is low, workload
Big problem, while also ensuring the accuracy of identification.
Detailed description of the invention
Attached drawing 1 is echinococcus protoscolex image.
Attached drawing 2 is the former ITTI notable figure that echinococcus protoscolex image is obtained by existing method.
Attached drawing 3 is to obtain improved ITTI notable figure after echinococcus protoscolex image is processed by the invention.
Attached drawing 4 is the recognition result figure after echinococcus protoscolex image detects through the invention.
Attached drawing 5 is the stream of the echinococcus protoscolex survival rate detection method of vision significance of the present invention and SIFT feature
Cheng Tu.
Specific embodiment
The present invention is not limited by the following examples, can determine according to the technique and scheme of the present invention with actual conditions specific
Embodiment.
The echinococcus protoscolex survival rate detection method of embodiment 1, the vision significance and SIFT feature, by following steps
Rapid to carry out: 20 to be detected to 100 echinococcus protoscolexs are refused the processing of dye method or Trypan Blue by Yihong by the first step
Processing, after processing and takes pictures, and obtains echinococcus protoscolex processing image to be detected;Second step, it is former to echinococcus to be detected
The polypide image saliency map of the head larva of a tapeworm or the cercaria of a schistosome processing image zooming-out color and brightness;Third step is significant to color and the polypide image of brightness
Figure carries out linear weighted function and generates total notable figure;4th step, extracts the marking area of total notable figure, finds in marking area doubtful
The central point of polypide simultaneously cuts all doubtful living worm body slices, comes out these suspected target zone markers in marking area,
Then by SIFT algorithm, the sift feature of living worm body slice is extracted, the sift feature vector of corresponding suspicious region is generated;The
Five steps, the sift after clustering in the sift feature vector of suspicious region and echinococcus protoscolex living worm body image data figure is special
Identification is compared in sign vector, if matching identification the result is that living worm body, labeled as living worm body target, if matching identification
Result be not living worm body, then cancel label, label result be finally restored to echinococcus protoscolex to be detected processing image
It goes up and counts, obtain echinococcus protoscolex survival rate.The echinococcus protoscolex of vision significance of the present invention and SIFT feature
The operation of survival rate detection method can carry out on computer software Matlab R2016a.
Embodiment 2, as the optimization of above-described embodiment, echinococcus protoscolex living worm body image data figure is in the steps below
Obtain: the first step takes 20 to 100 echinococcus protoscolexs to refuse the processing of dye method or Trypan Blue processing, place by Yihong
It after reason and takes pictures, obtains echinococcus protoscolex processing image;Second step is chosen 50 width to 70 width echinococcus protoscolexs and is handled
The background image of living worm body image and corresponding living worm body image in image establishes database;Third step is mentioned by SIFT algorithm
Take the sift feature vector of living worm body image and background image in database;4th step passes through k-means to sift feature vector
Algorithm is clustered, and then the sift feature vector after cluster is put into svm classifier, obtains echinococcus protoscolex worm living
Volumetric image data figure.It is existing known that the processing of dye method, Trypan Blue processing, SIFT algorithm and k-means algorithm are refused in Yihong
It is public.
Theoretical and algorithm (is referred to as protoscolex with echinococcus protoscolex in algorithm theoretical)
Vision significance
The vision system of the mankind possesses the ability of image understanding, identification, processing, with computer analog vision system, establishes vision
Attention model is the research hotspot in field of image processing.The visual observation of the mankind has selectivity, will not see sight
All things are all analyzed and are thought deeply, and brain is only concerned him and those of is concerned about things.In brief, brain is only handled obviously simultaneously
Special things, as significant things.The visual attention model emphasis of computer simulation be exactly quickly locate it is significant
Region.In a larger sense, image contains a variety of information that can be perceived by the mankind, such as: color, texture, brightness
Deng.But, it is generally the case that not all information is all that we are of concern, often the information of only a certain partial region
It is only interested to us.Therefore, area-of-interest is positioned, and it is extracted from image, just becomes image procossing
In unusual necessary step;Currently, the research of vision significance mainly has 4 kinds of models: spectrum residue model, Hu-Rajan-
Chia model, Stentiford model, ITTI visual attention model;Feature between Itti model foundation target and background,
Difference between contrast simulates human perception ability, extracts area-of-interest.
Improved ITTI model
Traditional to refuse the processing of dye method by Yihong due to the micro- protoscolex image analysis of medicine, color of image more divides after dyeing
It is bright, and ITTI vision mode mainly utilizes color characteristic, direction character, brightness, therefore ITTI vision mode is more to close
Suitable selection;The accuracy that the present invention extracts saliency region for tradition ITTI vision mode is not high, it is difficult to extract
The problem of whole area-of-interest, the fact that had differences according to human eye to different significant characteristics sensitivitys, to tradition
Significant characteristics combination improves in ITTI vision mode, and in protoscolex image, the direction of protoscolex does not have substantially
Play the role of too big, ignore direction notable figure, color and the notable figure of brightness are subjected to linear weighted function and generate total notable figure S;
After improving, obtained total notable figure S preferably can carry out coarse positioning to protoscolex.
Algorithm
SIFT algorithm finds extreme point in scale space, extracts position, scale, rotational invariants, to rotation, scaling, bright
Degree variation maintains the invariance, and can also keep stability to a certain extent to affine transformation, noise;It mainly includes following four
A step: the detection of scale space extreme point, characteristic point are filtered and are accurately positioned, characteristic point direction is distributed, key point description
Symbol;Just obtained complete characteristic point by three above step, each characteristic point include coordinate position, locating scale and
These three information of direction.By above step, for each key point, gather around there are three information: position, scale and direction.
Next be exactly to establish a descriptor for each key point, with one group of vector by this key point be depicted come, make its not with
Various change and change, such as illumination variation, visual angle change etc..This description not only includes key point, also comprising key
To its contributive pixel around point, and descriptor should have higher uniqueness, in order to improve characteristic point correct
The probability matched.When specific operation, usually key point and surrounding point are compared, centered on each extreme point
Point selects its 16 neighbouring sub-regions, sorts according to coordinate bit, can structure due to containing 8 directions in each region
Make the Feature Descriptor for generating 128 dimensions of the extreme point.It can be obtained scale invariant feature description of target image whole.
For the stability for further increasing Feature Descriptor, also it is wanted to do normalized.
View-based access control model conspicuousness and SIFT identification protoscolex algorithm are summarized
The present invention extracts refuse protoscolex image that dye method obtains by Yihong first by improved ITTI model and SIFT algorithm
Notable figure, the central point for obtaining doubtful protoscolex living cut institute in the position of central point according to the work protoscolex size to be detected
There is doubtful protoscolex image living.By SIFT algorithm, SIFT feature first is extracted to known protoscolex sample image living, is obtained
SIFT feature vector clusters vector with k-means clustering method, the feature vector of k dimension is generated, by this feature vector
SVM training is carried out, SVM classifier is obtained;Doubtful protoscolex image zooming-out SIFT feature living is put into trained svm classifier again
Device carries out "Yes" "No" two and classifies, then classification results are returned to original image and are marked, and just obtains final testing result.
The verification test (echinococcus protoscolex is referred to as protoscolex in verification test) of the above embodiment of the present invention
It in order to verify the validity of algorithm proposed in this paper, is tested at Matlab R2016a, the picture of use is by quilt
The processed protoscolex micro-image of dye method is refused in Yihong, and protoscolex samples pictures living are obtained not from different protoscolex pictures
Same target image and background image.60 width protoscolex image library living and 60 width background image libraries are initially set up, svm classifier is generated
Device, then through the invention the echinococcus protoscolex survival rate detection method of vision significance and SIFT feature to be detected
Protoscolex image is identified;It extracts image saliency map and marks doubtful protoscolex living;Doubtful protoscolex living is extracted
SIFT feature is put into trained SVM classifier classification, then classification results back to original protoscolex image and are marked
Out;Attached drawing 1 is echinococcus protoscolex image;Attached drawing 2 is the original that echinococcus protoscolex image is obtained by existing method
ITTI notable figure;Attached drawing 3 is to obtain improved ITTI notable figure after echinococcus protoscolex image is processed by the invention;Attached drawing 4
Recognition result figure after being detected through the invention for echinococcus protoscolex image;Attached drawing 5 be vision significance of the present invention and
The flow chart of the echinococcus protoscolex survival rate detection method of SIFT feature.Attached drawing 2 and attached drawing 3 can be seen that after comparing
Improved ITTI notable figure is compared with former ITTI notable figure, although omitting the direction of image, does not influence recognition effect, instead
Keep living worm body connected domain clearer, more effectively helps the rough identification to living worm body.
Test 1. takes to be refused in the processing of dye method or the processed protoscolex displaing micro picture to be detected of Trypan Blue altogether by Yihong
There are 23 complete protoscolexs, one of protoscolex is dead protoscolex, through the invention the spine of vision significance and SIFT feature
Ball tapeworm protoscolex survival rate detection method identifies that recognition result is shown in Table 1 to protoscolex image to be detected, table 1
In improved ITTI notable figure be vision significance of the present invention and SIFT feature echinococcus protoscolex survival rate detection method
Third step in, total notable figure of linear weighted function generation, 1 Central Plains ITTI of table are carried out to the polypide image saliency map of color and brightness
Notable figure is to carry out the total significant of linear weighted function generation to color, direction and the polypide of brightness image saliency map by existing method
Figure;As it can be seen from table 1 having identified 21 protoscolexs living altogether as the result is shown, discrimination reaches 95.4%, meets experiment
It is required that;Illustrate the echinococcus protoscolex survival rate detection method of vision significance through the invention and SIFT feature to be detected
The discrimination that is identified of protoscolex image reached 95.4%, higher than the discrimination of former ITTI notable figure, improved ITTI is aobvious
Work figure can more show the feature of echinococcosis polypide image compared with former ITTI notable figure, can accurately mark it is all can
Doubt target;Therefore, improved ITTI notable figure is better than the result that original ITTI notable figure obtains in the present invention.
Test 2. take by Yihong refuse the processing of dye method or the processed protoscolex displaing micro picture a to be detected of Trypan Blue, to
Protoscolex displaing micro picture b and protoscolex displaing micro picture c to be detected is detected, through the invention the spine of vision significance and SIFT feature
Ball tapeworm protoscolex survival rate detection method identifies protoscolex image to be detected, in protoscolex displaing micro picture to be detected
The experimental result of influence of the living worm body number to experimental result is as shown in table 2, from table 2 it can be seen that vision is aobvious through the invention
The echinococcus protoscolex survival rate detection method of work property and SIFT feature identifies protoscolex image to be detected, identifies
Rate is high, and discrimination 90% or more, the living worm body under microscope in polypide image can be identified substantially, thus for inspection
It tests drug effect and strong evidence is provided, while greatly reducing the workload of testing staff, avoid and led by existing artificial counting
Low efficiency, heavy workload and the error-prone problem of cause.
In conclusion the echinococcus protoscolex survival rate detection method of vision significance of the present invention and SIFT feature is treated
The protoscolex image of detection is identified that protoscolex living substantially falls in labeled image-region;View-based access control model conspicuousness
Recognition methods with SIFT feature is calculated image local, only considers the doubtful protoscolex region obtained by notable figure,
The plenty of time is saved from SIFT feature extraction, in the acquisition of the SIFT feature vector of sample image, the present invention uses k-
Means algorithm carries out cluster to it and changes SIFT feature vector there are the description of the feature of higher-dimension, and it is excessively multiple to solve calculating
It is miscellaneous, the problems such as taking long time, the computational efficiency that target search is carried out using SIFT feature is improved, in marking area
SIFT feature is more stable, changes the human error as caused by artificial counting, and the artificial counting time is longer, caused by work
The problem of low efficiency, heavy workload, while also ensuring the accuracy of identification.
The above technical features constitute embodiments of the present invention, can basis with stronger adaptability and implementation result
Actual needs increases and decreases non-essential technical characteristic, to meet the needs of different situations.
Claims (2)
1. the echinococcus protoscolex survival rate detection method of a kind of vision significance and SIFT feature, it is characterised in that by following
Step carries out: the first step, and 20 to be detected to 100 echinococcus protoscolexs are refused the processing of dye method by Yihong or trypan blue contaminates
Color processing, after processing and takes pictures, and obtains echinococcus protoscolex processing image to be detected;Second step, to echinococcus to be detected
The polypide image saliency map of protoscolex processing image zooming-out color and brightness;Third step, it is aobvious to the polypide image of color and brightness
It writes figure and carries out the total notable figure of linear weighted function generation;4th step, extracts the marking area of total notable figure, finds in marking area doubtful
Like polypide central point and cut all doubtful living worm bodies slices, these suspected target zone markers are gone out in marking area
Come, then by SIFT algorithm, extract the sift feature of living worm body slice, generate the sift feature of corresponding suspicious region to
Amount;5th step, after being clustered in the sift feature vector of suspicious region and echinococcus protoscolex living worm body image data figure
Identification is compared in sift feature vector, if matching identification the result is that living worm body, be labeled as living worm body target, if than
Result to identification is not living worm body, then cancels label, and finally label result is restored at echinococcus protoscolex to be detected
It on reason image and counts, obtains echinococcus protoscolex survival rate.
2. the echinococcus protoscolex survival rate detection method of vision significance according to claim 1 and SIFT feature,
It is characterized in that echinococcus protoscolex living worm body image data figure obtains in the steps below: the first step takes 20 to 100 spine
Ball tapeworm protoscolex refuses the processing of dye method by Yihong or Trypan Blue is handled, and after processing and takes pictures, obtains echinococcus procephalon
The larva of a tapeworm or the cercaria of a schistosome handles image;Second step chooses living worm body image in 50 width to 70 width echinococcus protoscolexs processing image and to taking on service jobs
The background image of polypide image establishes database;Third step extracts living worm body image and background in database by SIFT algorithm
The sift feature vector of image;4th step clusters sift feature vector by k-means algorithm, after then clustering
Sift feature vector be put into svm classifier, obtain echinococcus protoscolex living worm body image data figure.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810476337.8A CN108960042B (en) | 2018-05-17 | 2018-05-17 | Echinococcus proctostermias survival rate detection method based on visual saliency and SIFT characteristics |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810476337.8A CN108960042B (en) | 2018-05-17 | 2018-05-17 | Echinococcus proctostermias survival rate detection method based on visual saliency and SIFT characteristics |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108960042A true CN108960042A (en) | 2018-12-07 |
CN108960042B CN108960042B (en) | 2021-06-08 |
Family
ID=64499811
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810476337.8A Active CN108960042B (en) | 2018-05-17 | 2018-05-17 | Echinococcus proctostermias survival rate detection method based on visual saliency and SIFT characteristics |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108960042B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111275717A (en) * | 2020-01-17 | 2020-06-12 | 杭州迪英加科技有限公司 | Contrast display method for different stained sections, electronic device and readable medium |
CN111429412A (en) * | 2020-03-17 | 2020-07-17 | 北京青燕祥云科技有限公司 | Ultrasound AI auxiliary diagnosis method and system for hydatid hepatica |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102360435A (en) * | 2011-10-26 | 2012-02-22 | 西安电子科技大学 | Undesirable image detecting method based on connotative theme analysis |
US20160104055A1 (en) * | 2014-10-09 | 2016-04-14 | Adobe Systems Incorporated | Image Cropping Suggestion Using Multiple Saliency Maps |
CN107357834A (en) * | 2017-06-22 | 2017-11-17 | 浙江工业大学 | A kind of image search method of view-based access control model conspicuousness fusion |
CN107704864A (en) * | 2016-07-11 | 2018-02-16 | 大连海事大学 | Well-marked target detection method based on image object Semantic detection |
-
2018
- 2018-05-17 CN CN201810476337.8A patent/CN108960042B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102360435A (en) * | 2011-10-26 | 2012-02-22 | 西安电子科技大学 | Undesirable image detecting method based on connotative theme analysis |
US20160104055A1 (en) * | 2014-10-09 | 2016-04-14 | Adobe Systems Incorporated | Image Cropping Suggestion Using Multiple Saliency Maps |
CN107704864A (en) * | 2016-07-11 | 2018-02-16 | 大连海事大学 | Well-marked target detection method based on image object Semantic detection |
CN107357834A (en) * | 2017-06-22 | 2017-11-17 | 浙江工业大学 | A kind of image search method of view-based access control model conspicuousness fusion |
Non-Patent Citations (2)
Title |
---|
SHAO H 等: "Retrieving images using saliency detection and graph matching", 《IEEE》 * |
若山古丽•肉孜 等: "刃天青还原法检测细粒棘球绦虫原头蚴总数", 《中国人兽共患病学报》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111275717A (en) * | 2020-01-17 | 2020-06-12 | 杭州迪英加科技有限公司 | Contrast display method for different stained sections, electronic device and readable medium |
CN111275717B (en) * | 2020-01-17 | 2023-04-11 | 杭州迪英加科技有限公司 | Contrast display method for different stained sections, electronic device and readable medium |
CN111429412A (en) * | 2020-03-17 | 2020-07-17 | 北京青燕祥云科技有限公司 | Ultrasound AI auxiliary diagnosis method and system for hydatid hepatica |
CN111429412B (en) * | 2020-03-17 | 2023-11-03 | 北京青燕祥云科技有限公司 | Ultrasonic AI auxiliary diagnosis method and system for liver bag worm |
Also Published As
Publication number | Publication date |
---|---|
CN108960042B (en) | 2021-06-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112070772B (en) | Blood leukocyte image segmentation method based on UNet++ and ResNet | |
Gao et al. | Automatic change detection in synthetic aperture radar images based on PCANet | |
Jahanifar et al. | Supervised saliency map driven segmentation of lesions in dermoscopic images | |
Wang et al. | Using deep convolutional neural networks for multi-classification of thyroid tumor by histopathology: a large-scale pilot study | |
Spanhol et al. | A dataset for breast cancer histopathological image classification | |
CN106529424B (en) | A kind of logo detection recognition method and system based on selective search algorithm | |
Alexe et al. | What is an object? | |
ES2641480T3 (en) | Machine learning system based on tissue objects for automated slide scoring full digital | |
CN110348319A (en) | A kind of face method for anti-counterfeit merged based on face depth information and edge image | |
JP5394485B2 (en) | Signet ring cell detector and related methods | |
US20150125052A1 (en) | Drusen lesion image detection system | |
CN104392240A (en) | Parasite egg identification method based on multi-feature fusion | |
Pan et al. | Mitosis detection techniques in H&E stained breast cancer pathological images: A comprehensive review | |
CN107066972B (en) | Natural scene Method for text detection based on multichannel extremal region | |
CN107679509A (en) | A kind of small ring algae recognition methods and device | |
Yanikoglu et al. | Sabanci-Okan System at ImageClef 2012: Combining Features and Classifiers for Plant Identification. | |
Yang et al. | Pathminer: a web-based tool for computer-assisted diagnostics in pathology | |
WO2014066231A1 (en) | Cell identification method and device, and urine analyzer | |
CN108550148A (en) | Nucleus in histotomy micro-image divides automatically and classifying identification method | |
Meng et al. | Text detection in natural scenes with salient region | |
CN110264454A (en) | Cervical cancer tissues pathological image diagnostic method based on more hidden layer condition random fields | |
CN108960042A (en) | The echinococcus protoscolex survival rate detection method of vision significance and SIFT feature | |
Khan et al. | Comparitive study of tree counting algorithms in dense and sparse vegetative regions | |
Kuruvilla et al. | Otitis media vocabulary and grammar | |
San et al. | Constrained-MSER detection of retinal pathology |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |