CN106096255A - A kind of infrared diagnostics system and method - Google Patents

A kind of infrared diagnostics system and method Download PDF

Info

Publication number
CN106096255A
CN106096255A CN201610395354.XA CN201610395354A CN106096255A CN 106096255 A CN106096255 A CN 106096255A CN 201610395354 A CN201610395354 A CN 201610395354A CN 106096255 A CN106096255 A CN 106096255A
Authority
CN
China
Prior art keywords
retrieval
module
infrared
image
coloured image
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201610395354.XA
Other languages
Chinese (zh)
Inventor
田思
徐璟
李永平
章晓敏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ningbo Dahongying University
Original Assignee
Ningbo Dahongying University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Ningbo Dahongying University filed Critical Ningbo Dahongying University
Priority to CN201610395354.XA priority Critical patent/CN106096255A/en
Publication of CN106096255A publication Critical patent/CN106096255A/en
Pending legal-status Critical Current

Links

Classifications

    • G06F19/321—
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems

Landscapes

  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Medical Informatics (AREA)
  • Biomedical Technology (AREA)
  • Public Health (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Pathology (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Image Analysis (AREA)

Abstract

The present invention relates to field of medical technology, particularly relate to a kind of infrared diagnostics system and method.Coloured image is diagnosed by the present invention by the retrieval information matching with coloured image mates most, the retrieval information can mated most by this, i.e. diagnoses disease present in coloured image, reduces the requirement of diagnosis.

Description

A kind of infrared diagnostics system and method
Technical field
The present invention relates to field of medical technology, particularly relate to a kind of infrared diagnostics system and method.
Background technology
Current infrared diagnostics has medically been applied, but needs in shooting infrared during existing infrared diagnostics After image, artificial is further diagnosed, and during Artificial Diagnosis, has higher requirement, institute to diagnosis capability With, infrared diagnosis technology owing to requiring higher not to be widely used.
Content of the invention
The problem existing for prior art, now provides a kind of infrared diagnostics system and method.
Concrete technical scheme is as follows:
A kind of infrared diagnostics system, comprising:
Infrared photography module, utilizes infrared imagery technique to obtain gray level image；
Coding module, is connected with described infrared photography module, and described gray level image carries out color processing, and output one is colored Image, to show infrared information by described coloured image；
Characteristic extracting module, is connected with described coding module, carries out feature extraction to described coloured image；
Retrieval module, is connected with described characteristic extracting module, is pre-stored with multiple standard information, profit in described retrieval module Select the retrieval information with described coloured image adaptation by the feature extracted in described standard information；
Matching module, with described retrieval module is connected, utilize preset algorithm be calculated in described retrieval information one and The retrieval information that described coloured image mates most.
Preferably, described characteristic extracting module includes:
Segmentation module, is multiple regions by described color images；
Extraction module, is connected with described segmentation module, extracts region interested in the plurality of region；
Build module, be connected with described extraction module, feature extraction carried out to described region interested, with construction feature Space.
Preferably, the feature that described characteristic extracting module is extracted includes: color, shape and texture.
Preferably, the search method of described retrieval module includes: based on the retrieval of color characteristic, based on the inspection of shape facility Rope and the retrieval based on textural characteristics.
Preferably, described preset algorithm is genetic algorithm.
A kind of infrared diagnostics method, comprising:
Step S1, utilizes infrared imagery technique to obtain gray level image；
Described gray level image is carried out color processing by step S2, exports a coloured image, to be shown by described coloured image Show infrared information；
Step S3, carries out feature extraction to described coloured image；
Step S4, utilizes the feature extracted to select the inspection with described coloured image adaptation in the standard information of pre-stored Rope information；
Step S5, utilizes preset algorithm to be calculated an inspection mated most with described coloured image in described retrieval information Rope information.
Preferably, described step S3 includes:
Step S31, is multiple regions by described color images；
Step S32, extracts region interested in the plurality of region；
Step S33, carries out feature extraction to described region interested, with construction feature space.
Preferably, the feature of extraction includes: color, shape and texture.
Preferably, the search method in described step S4 includes: based on the retrieval of color characteristic, based on the inspection of shape facility Rope and the retrieval based on textural characteristics.
Preferably, described preset algorithm is genetic algorithm.
Technique scheme provides the benefit that:
In technique scheme, by the retrieval information matching with coloured image mates most, can be mated most by this Retrieval information coloured image is diagnosed, i.e. disease present in coloured image is diagnosed, reduces wanting of diagnosis Ask.
Brief description
Fig. 1 is the structural representation of the embodiment of the present invention a kind of infrared diagnostics system；
Fig. 2 is the flow chart of the embodiment of the present invention a kind of infrared diagnostics method.
Detailed description of the invention
It should be noted that in the case of not conflicting, following technical proposals, can be mutually combined between technical characteristic.
Below in conjunction with the accompanying drawings the detailed description of the invention of the present invention is further described:
A kind of infrared diagnostics system, as shown in Figure 1, comprising:
Infrared photography module, utilizes infrared imagery technique to obtain gray level image；
Coding module, is connected with infrared photography module, and gray level image is carried out color processing, exports a coloured image, with Show infrared information by coloured image；
Characteristic extracting module, is connected with coding module, carries out feature extraction to coloured image；
Retrieval module, is connected with characteristic extracting module, is pre-stored with multiple standard information, utilizes extraction in retrieval module Feature selects the retrieval information with coloured image adaptation in standard information；
Matching module, is connected with retrieval module, utilizes preset algorithm to be calculated one and coloured image in retrieval information The retrieval information mated most.
In the present embodiment, infrared photography module is utilized related infrared imagery technique to generate by thermal infrared imager, infrared one-tenth As technology is a kind of radiation information Detection Techniques, utilizes the electronic installation that certain is special, the heat distribution of object is converted into gray scale Image, and with gray level or pseudo-color processing out, thus obtain the Temperature Distribution field of measured target.Infrared imagery technique one As be to passively receive the infra-red radiation of object and imaging, characterize the thermo parameters method of object by the gray scale of image.Due to thing The reason that body heat balance and according to Planck law, the image local using infrared imagery technique to shoot is adjacent has higher sky Between Gray Correlation.
In the present embodiment, owing to the color sensitive cell of human eye can tell thousand of kinds of color shadeses and brightness, but right Black and white gray level is insensitive.The infrared image that infrared photography module produces is black and white grayscale image (gray level image), gray scale Value dynamic range is little.Therefore, human eye is difficult to obtain abundant information from these gray levels.In order to more intuitively strengthen display The level of image, improves human eye resolution capability, and coding module carries out Pseudo Col ored Image to gray level image, thus reaches image enhaucament Effect, make image information abundanter.On the basis of rainbow encodes, in conjunction with medical infrared thermography system, characteristics of image is examined The requirement surveyed and the algorithm research that pseudo-color coding is processed so that different images and different regions are because of the difference of temperature Show different colors, and coloured image is well arranged, clear color.
In the present embodiment, characteristic extracting module is realized by computer vision technique and image processing techniques, refers to It is the information using computer to extract in coloured image, determine whether the point of each coloured image belongs to a characteristics of image.Special The result levying extraction is that the point on coloured image is divided into different subsets, and these subsets tend to belong to point, the continuous print isolating Curve or continuous print region.
The retrieval module of the present embodiment is that CBIR technology carries out image retrieval, based on the image of content Retrieval technique refers to directly enter line retrieval according to the various features describing media object picture material, and it can be from the standard of database Finding in information and having specific characteristic or the retrieval information (including video segment) containing certain content, it is different from traditional Based on the retrieval method of keyword, merge the technology such as image understanding, pattern-recognition.
The framework knot of the image database system of an information retrieval based on contents is realized under wide area network (Internet) environment Structure can be divided into two parts: the establishment of characteristics of image index and image retrieval.In general, the analysis of characteristics of image and index Establishment completes at server end off-line, its major function include image warehouse-in before pretreatment, the content characteristic of image Extract the coding describing with analysis, characteristics of image and storage.For the retrieval of image, its main task is that user is interested Image instance submit to server, the feature of its content is extracted and represents.Then image retrieval engine is called by one Fixed similar to search method carries out the Similarity Measure between image, and the similar image (retrieval information) obtaining inquiry is by them Similarity distance be ranked up from small to large, and by result feed back.
In one preferred embodiment of the present invention, characteristic extracting module includes:
Segmentation module, is multiple regions by color images；
Extraction module, is connected with segmentation module, extracts region interested in multiple regions；
Build module, be connected with extraction module, feature extraction is carried out to region interested, with construction feature space.
In the present embodiment, first coloured image was split before feature extraction, extract area-of-interest and be related to Process.Color images is the basic work of image procossing and computer vision field.The quantity of image partition method is non- Chang Duo, the present embodiment is illustrated with energy minimization method, and the basic step of energy minimization method is: 1. design one Individual object function (energy function), the corresponding optimal solution of its minimum of a value, two conventional constraints are data and priori.Data are about Bundle limits ideal solution should be as far as possible close with True Data；The prior-constrained form requiring ideal solution should be protected with priori Hold consistent；2. minimizing object function, most of energy functions interested are non-convex, have multiple minimum, cause majority Method is only able to find Approximating Solutions, therefore, minimizes process generally relatively difficult.The constraint of suitable encoded question forms energy letter Count and find a good minimum method to be of equal importance, because any one walks out of the wrong inefficacy that all can cause segmentation.
Energy minimization method, gives the Unified frame of segmentation, can solve by standard optimization techniques, 2. provides bright True thinking encodes the constraint of problem to be solved, and is equivalent owing to minimizing the MAP estimation of energy function and MRF, Also the correctness of energy minimization method can be solved from Bayesian statistics angle.Energy minimization method has multiple classification side Method, including global and local method, certainty and randomization method, continuity and discreteness method etc..Frequently with overall situation drawn game The sorting technique in portion, common global energy minimization method includes simulated annealing method, dynamic programming method and Graph-theoretical Approach Deng；Local energy minimizes method and includes variational method and ICM (Iterated conditional modes) method etc..Simulation Method for annealing represents a class randomized optimization process；Dynamic Programming is a multistep compound theory, by a N step process is turned Turn to N number of single step process and realize being converted into global optimum local optimum sum；The main thought of Graph-theoretical Approach is to reflect image Penetrating into weighted graph, image pixel being seen the summit of mapping, the relation between adjacent pixels sees the limit of mapping, between adjacent pixels Similitude regards the weights on limit as, the weights design energy function according to limit, completes the segmentation to figure by minimizing energy function, Thus realize that image is split；Variational method is geometric active contour model (Geometry active contour model) Energy minimization method；ICM method is based on a determination that property greedy strategy carrys out the energy minimization method of optimizing.
In above-described embodiment, the feature extraction that is related to for segmentation result, first application normalization segmentation (Normalized cut, Ncut) method carries out the segmentation of image, extracts area-of-interest；Then to the area-of-interest extracting Carry out feature extraction constitutive characteristic space, then diagnosing by retrieval module and matching module, drawing diagnostic result.
Ncut criterion is a kind of without supervision image Segmentation Technology, and it does not needs to initialize, and has 3 main features: (1) it divides the image into the partition problem that problem is converted to figure；(2) it is an overall situation criterion；(3) it maximizes difference simultaneously Dissimilarity between group and the similitude in same group.The general principle of Ncut criterion is: for a given figure G= (V, E), it is assumed that be classified as two disjoint parts A and B, A ∪ B=V, the dissimilar degree of the two part can be defined as The synthesis of the power on original all limits connecting two parts and being left out now:
c u t ( A , B ) = Σ i ∈ A , j ∈ B w ( i , j )
W (i, j) the i.e. power on the limit of tie point i and some j, the similarity degree between its expression 2 in formula.One width figure is Excellent dichotomy is i.e. the value minimum making cut, but directly proportional to the number cutting middle limit owing to cutting, and therefore minimal cut is generally not It is exactly that optimum cuts.
In one preferred embodiment of the present invention, the feature that characteristic extracting module is extracted includes: color, shape and texture.
In one preferred embodiment of the present invention, the search method of retrieval module includes: based on the retrieval of color characteristic, base Retrieval in shape facility and the retrieval based on textural characteristics.
Three kinds of retrieval modes in above-described embodiment particularly as follows:
First, the retrieval based on color characteristic: the color of coloured image is a kind of visual characteristic of body surface, every kind of thing Body has its distinctive color characteristic, as people mention green related with trees or grassland often, speak of blue be often with Sea or blue sky are related to.Same class object often has similar color characteristic, therefore can distinguish thing according to color characteristic Body, the present embodiment uses histogram to describe color characteristic.
Secondly, based on the retrieval of shape facility: shape facility is one of key element describing coloured image, and it can be relatively Reflecting well the feature in region, therefore this feature is an important available factor in CBIR technology.At X-Y scheme In image space, shape is typically considered a contour curve area encompassed closed, so the description to shape relates to To the description on profile border and the description to this border institute enclosing region.Current encloses mostly based on Similar Shape Retrieval Set up image index around from the contour feature of shape and the provincial characteristics of shape.Shape include area, connectedness, circularity, The feature such as eccentricity, major axes orientation.
Finally, the retrieval based on textural characteristics: texture refers to certain regular change of image pixel gray level collection or color. Textural characteristics mainly includes roughness, directionality, contrast and systematicness.Texture based image retrieval generally uses statistics side Method, structural approach and frequency spectrum analysis method are carried out.
In one preferred embodiment of the present invention, preset algorithm is genetic algorithm.
In the present embodiment, matching module is substantially carried out image steganalysis and pattern match, the general principle of pattern-recognition It is, by an input pattern compared with the multiple mode standards preserving in systems, to find out closest mode standard, should Class name representated by mode standard exports as the class name of input pattern.The image information of acquisition is being carried by this research through feature Using algorithm for pattern recognition to carry out pattern match through image retrieval after taking, final coupling obtains inside corresponding expert system Pre-stored information.
The application uses genetic algorithm (Genetic Algorithm, be called for short GA) to realize the pattern-recognition of image and mate. Genetic algorithm is the theory of natural selection and Population Genetics principle based on Darwinian evolutionism, uses for reference in nature natural Selection, the rule of the survival of the fittest, use a kind of algorithm that the corresponding name in science of heredity and method are set up in a computer.Heredity Algorithm is set up on the basis of natural selection and population genetic, the evolutionary process of simulation nature " survival of the fittest in natural selection, the survival of the fittest ", Problem space carries out parallel, the random chess game optimization of the overall situation so that population becomes the convergence of global optimum.
In GA enforcement, 3 of most critical are: define coding structure, determine fitness function and definition genetic operator.Utilize Markov theory, carries out Analysis On The Global Convergence to the genetic algorithm using actual coding, has shown that at colony's number be infinity When can converge to the conclusion of globally optimal solution.Although it is set up on the basis of population size is infinitely great, but has in limited time in scale, receive Hold back the overall situation more advantage can reach.Based on this, when carrying out genetic Algorithm Design, individual chromosome bit string is by grader The cascade of all real number weights is constituted, each chromosome be represented by (W1T, W2T, W3T ..., W1T) the real vector of type, wherein Wi (i=1,2 ..., M) it is the weight vector of grader.So, each individuality in colony is exactly a grader.Grader The differentiation rate concentrated in test sample as the adaptive value of individual chromosome, wherein use Yik (i=1,2 ... N) represent kth time During iteration, i-th in colony is individual.
A kind of infrared diagnostics method, as shown in Figure 2, comprising:
Step S1, utilizes infrared imagery technique to obtain gray level image；
Gray level image is carried out color processing by step S2, exports a coloured image, to show infrared letter by coloured image Breath；
Step S3, carries out feature extraction to coloured image；
Step S4, utilizes the feature extracted to select the retrieval letter with coloured image adaptation in the standard information of pre-stored Breath；
Step S5, utilize preset algorithm be calculated in retrieval information one with the retrieval information that coloured image mates most.
In one preferred embodiment of the present invention, step S3 includes:
Step S31, is multiple regions by color images；
Step S32, extracts region interested in multiple regions；
Step S33, carries out feature extraction to region interested, with construction feature space.
In one preferred embodiment of the present invention, the feature of extraction includes: color, shape and texture.
In one preferred embodiment of the present invention, the search method in step S4 includes: based on the retrieval of color characteristic, base Retrieval in shape facility and the retrieval based on textural characteristics.
In one preferred embodiment of the present invention, preset algorithm is genetic algorithm.
In above-described embodiment, by characteristic extracting module by feature extraction, and image retrieval module carries out coloured image Retrieval, output one retrieval information, retrieval information and matching module interact, so that it is determined that the retrieval information mated most, Coloured image is diagnosed by the retrieval information can mated most by this, i.e. examines disease present in coloured image Disconnected, reduce the requirement of diagnosis.
By explanation and accompanying drawing, give the exemplary embodiments of the ad hoc structure of detailed description of the invention, based on present invention essence God, also can make other conversion.Although foregoing invention proposes existing preferred embodiment, but, these contents are not intended as Limitation.
For a person skilled in the art, after reading described above, various changes and modifications will be apparent to undoubtedly. Therefore, appending claims should regard whole variations and modifications of true intention and the scope covering the present invention as.In power The scope of any and all equivalence and content in the range of profit claim, be all considered as still belonging to the intent and scope of the invention.

Claims (10)

1. an infrared diagnostics system, it is characterised in that include:
Infrared photography module, utilizes infrared imagery technique to obtain gray level image；
Coding module, is connected with described infrared photography module, described gray level image is carried out color processing, export a cromogram Picture, to show infrared information by described coloured image；
Characteristic extracting module, is connected with described coding module, carries out feature extraction to described coloured image；
Retrieval module, is connected with described characteristic extracting module, is pre-stored with multiple standard information in described retrieval module, and utilization carries The feature taking selects the retrieval information with described coloured image adaptation in described standard information；
Matching module, with described retrieval module be connected, utilize preset algorithm be calculated in described retrieval information one with described The retrieval information that coloured image mates most.
2. infrared diagnostics system according to claim 1, it is characterised in that described characteristic extracting module includes:
Segmentation module, is multiple regions by described color images；
Extraction module, is connected with described segmentation module, extracts region interested in the plurality of region；
Build module, be connected with described extraction module, feature extraction is carried out to described region interested, empty with construction feature Between.
3. infrared diagnostics system according to claim 1, it is characterised in that the feature bag that described characteristic extracting module is extracted Include: color, shape and texture.
4. infrared diagnostics system according to claim 3, it is characterised in that the search method of described retrieval module includes: Based on the retrieval of color characteristic, the retrieval based on shape facility and the retrieval based on textural characteristics.
5. infrared diagnostics system according to claim 1, it is characterised in that described preset algorithm is genetic algorithm.
6. an infrared diagnostics method, it is characterised in that include:
Step S1, utilizes infrared imagery technique to obtain gray level image；
Described gray level image is carried out color processing by step S2, exports a coloured image, to show red by described coloured image External information；
Step S3, carries out feature extraction to described coloured image；
Step S4, utilizes the feature extracted to select the retrieval letter with described coloured image adaptation in the standard information of pre-stored Breath；
Step S5, utilizes preset algorithm to be calculated one in described retrieval information and believes with the retrieval that described coloured image mates most Breath.
7. infrared diagnostics method according to claim 1, it is characterised in that described step S3 includes:
Step S31, is multiple regions by described color images；
Step S32, extracts region interested in the plurality of region；
Step S33, carries out feature extraction to described region interested, with construction feature space.
8. infrared diagnostics method according to claim 1, it is characterised in that the feature of extraction includes: color, shape and line Reason.
9. infrared diagnostics method according to claim 8, it is characterised in that the search method in described step S4 includes: Based on the retrieval of color characteristic, the retrieval based on shape facility and the retrieval based on textural characteristics.
10. infrared diagnostics method according to claim 6, it is characterised in that described preset algorithm is genetic algorithm.
CN201610395354.XA 2016-06-03 2016-06-03 A kind of infrared diagnostics system and method Pending CN106096255A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610395354.XA CN106096255A (en) 2016-06-03 2016-06-03 A kind of infrared diagnostics system and method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610395354.XA CN106096255A (en) 2016-06-03 2016-06-03 A kind of infrared diagnostics system and method

Publications (1)

Publication Number Publication Date
CN106096255A true CN106096255A (en) 2016-11-09

Family

ID=57447094

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610395354.XA Pending CN106096255A (en) 2016-06-03 2016-06-03 A kind of infrared diagnostics system and method

Country Status (1)

Country Link
CN (1) CN106096255A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106778047A (en) * 2017-03-06 2017-05-31 武汉嫦娥医学抗衰机器人股份有限公司 A kind of traditional Chinese medical science facial diagnosis integrated system based on various dimensions medical image

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101639858A (en) * 2009-08-21 2010-02-03 深圳创维数字技术股份有限公司 Image search method based on target area matching
CN101789005A (en) * 2010-01-22 2010-07-28 深圳创维数字技术股份有限公司 Image searching method based on region of interest (ROI)
CN103207910A (en) * 2013-04-08 2013-07-17 河南大学 Image retrieval method based on hierarchical features and genetic programming relevance feedback
CN103217397A (en) * 2013-01-23 2013-07-24 山西省电力公司晋城供电分公司 SF6 gas detection method based on infrared image processing
CN103279938A (en) * 2013-04-03 2013-09-04 昆明物理研究所 Infrared/ glimmer image fusion night vision system
CN105335708A (en) * 2015-10-22 2016-02-17 四川膨旭科技有限公司 System for identifying road signs and guideboards during process of vehicle moving

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101639858A (en) * 2009-08-21 2010-02-03 深圳创维数字技术股份有限公司 Image search method based on target area matching
CN101789005A (en) * 2010-01-22 2010-07-28 深圳创维数字技术股份有限公司 Image searching method based on region of interest (ROI)
CN103217397A (en) * 2013-01-23 2013-07-24 山西省电力公司晋城供电分公司 SF6 gas detection method based on infrared image processing
CN103279938A (en) * 2013-04-03 2013-09-04 昆明物理研究所 Infrared/ glimmer image fusion night vision system
CN103207910A (en) * 2013-04-08 2013-07-17 河南大学 Image retrieval method based on hierarchical features and genetic programming relevance feedback
CN105335708A (en) * 2015-10-22 2016-02-17 四川膨旭科技有限公司 System for identifying road signs and guideboards during process of vehicle moving

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106778047A (en) * 2017-03-06 2017-05-31 武汉嫦娥医学抗衰机器人股份有限公司 A kind of traditional Chinese medical science facial diagnosis integrated system based on various dimensions medical image

Similar Documents

Publication Publication Date Title
CN111368896B (en) Hyperspectral remote sensing image classification method based on dense residual three-dimensional convolutional neural network
Zhang et al. A feature difference convolutional neural network-based change detection method
CN112380952B (en) Power equipment infrared image real-time detection and identification method based on artificial intelligence
CN109670528B (en) Data expansion method facing pedestrian re-identification task and based on paired sample random occlusion strategy
Santo et al. Deep photometric stereo network
CN107563355B (en) Hyperspectral anomaly detection method based on generation of countermeasure network
Chen et al. Hyperspectral image classification using dictionary-based sparse representation
CN114926746B (en) SAR image change detection method based on multiscale differential feature attention mechanism
Kang et al. Person re-identification between visible and thermal camera images based on deep residual CNN using single input
Xu et al. High-resolution remote sensing image change detection combined with pixel-level and object-level
CN111444960A (en) Skin disease image classification system based on multi-mode data input
CN106295124A (en) Utilize the method that multiple image detecting technique comprehensively analyzes gene polyadenylation signal figure likelihood probability amount
Miao et al. Classification of farmland images based on color features
Gu et al. Blind image quality assessment via vector regression and object oriented pooling
CN111680579B (en) Remote sensing image classification method for self-adaptive weight multi-view measurement learning
Chen et al. Agricultural remote sensing image cultivated land extraction technology based on deep learning
CN106960182A (en) A kind of pedestrian integrated based on multiple features recognition methods again
Yan et al. Monocular depth estimation with guidance of surface normal map
CN107798345A (en) Based on the diagonal EO-1 hyperion camouflaged target detection method with low-rank representation of block
CN112115806B (en) Remote sensing image scene accurate classification method based on Dual-ResNet small sample learning
CN109919246A (en) Pedestrian's recognition methods again based on self-adaptive features cluster and multiple risks fusion
Saikumar et al. Colour based image segmentation using fuzzy c-means clustering
CN109947960B (en) Face multi-attribute joint estimation model construction method based on depth convolution
CN114913409A (en) Camouflage target identification method for marine organisms
Ren et al. Multiscale adaptive convolution for hyperspectral image classification

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20161109

RJ01 Rejection of invention patent application after publication