CN104361135A - Image retrieval method - Google Patents
Image retrieval method Download PDFInfo
- Publication number
- CN104361135A CN104361135A CN201410752515.7A CN201410752515A CN104361135A CN 104361135 A CN104361135 A CN 104361135A CN 201410752515 A CN201410752515 A CN 201410752515A CN 104361135 A CN104361135 A CN 104361135A
- Authority
- CN
- China
- Prior art keywords
- image
- images
- difference
- similar
- classification
- 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
- 238000000034 method Methods 0.000 title claims abstract description 49
- 239000002245 particle Substances 0.000 claims abstract description 35
- 230000008569 process Effects 0.000 claims abstract description 21
- 238000005192 partition Methods 0.000 claims abstract description 10
- 230000008878 coupling Effects 0.000 claims description 6
- 238000010168 coupling process Methods 0.000 claims description 6
- 238000005859 coupling reaction Methods 0.000 claims description 6
- 230000006872 improvement Effects 0.000 claims description 6
- 230000011218 segmentation Effects 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 238000005457 optimization Methods 0.000 abstract description 6
- 230000000694 effects Effects 0.000 abstract 1
- 238000010845 search algorithm Methods 0.000 abstract 1
- 238000005516 engineering process Methods 0.000 description 4
- 230000001149 cognitive effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/758—Involving statistics of pixels or of feature values, e.g. histogram matching
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses an image retrieval method, which comprises the following concrete implementation processes: extracting gray histogram information from the template image and the source image according to a new partition idea; converting the problem of finding the most similar template images into classification through improved particle swarm optimization; continuously matching the images with large similarity according to the steps to obtain the most similar image; the proposal of an improved fast search algorithm during searching. Compared with the prior art, the image retrieval method can rapidly match similar image groups in the source image database, and greatly reduces the time and space complexity of an image retrieval algorithm and improves the image retrieval efficiency while keeping the excellent retrieval effect.
Description
Technical field
The present invention relates to technical field of information retrieval, specifically a kind of practical, based on the image search method of the fast finding algorithm of intelligent algorithm and improvement.
Background technology
The development of accompanying information technology, the application of image retrieval is more extensive, and become indispensable technology, the important indicator weighing image retrieval algorithm quality is exactly Time & Space Complexity in addition to accuracy.Image retrieval purposes in real world applications is very extensive.Its principle refers to the similar or identical image by algorithm search pattern image in image data base.Most crucial problem remains the eigenwert of comparison two width image, the model of problem: a given secondary size is respectively
image, source images:
, template image: in formula
for the gray-scale value of image.Traditional image retrieval algorithm calculated amount is large, precision is little.
At present, image retrieval algorithm has a variety of, performance varies, ubiquity recall precision is low, the problem that precision ratio and recall ratio are not high, based on this, now provide a kind of by population (Particle Swarm Optimization, PSO) improvement of algorithm and new classificating thought improved after image search method, the method is while maintenance retrieval effectiveness excellence, and what achieve the Time & Space Complexity of image retrieval algorithm significantly reduces the recall precision that improve image.
Summary of the invention
Technical assignment of the present invention is for above weak point, provides a kind of practical, image search method.
A kind of image search method, its specific implementation process is:
One, template image and source images are extracted grey level histogram information according to partition method, described partition method is specially:
Carrying out subregion to two width images gets a little, becomes the core of both comparisons after some images by every width Iamge Segmentation;
The difference threshold of two width image co-located region is set, then the difference between co-located region corresponding to two images is calculated clockwise, whether the size according to difference completes classification close to threshold value, is categorized as similar image close to threshold value, otherwise is classified as non-similar image;
Two, by find template image the most similar problem change into by improve particle group optimizing classify, the particle group optimizing process of this improvement is:
First the dynamic initialization of population is carried out, by the image difference of the given number of random selecting as population;
In image data base, retrieve after dynamic initialization, choose the image collection of convergence after learning certain number of times as the first kind;
Continue studying n time, the image collection choosing convergence is as Equations of The Second Kind, and with this recurrence until reach amount of images threshold value in classification, in optimizing process, particle is followed current location and velocity optimal particle in colony and moved, and through obtaining optimum solution by for after iterative search, finally return classification value;
Three, repeat step one, two, coupling draws the most similar image;
Four, image retrieval is completed by fast finding algorithm, the particular content of this fast finding algorithm is: by analyzing the difference of the histogram information of the partition sections of image and the template image retrieved, to the quicksort that difference portion carries out from small to large, former width images that value is minimum, carry out quicksort, when data length is greater than threshold value time, data being divided equally is two sections, quicksort is carried out, successively recurrence in each section; Finally return and come the most front image, be the image searched.
In described step one, the detailed process of Iamge Segmentation is:
Get the module that two, the centre of image is wide, namely width is 1/16th of whole image;
The mid line region that the centre of getting every a 1/4th images interlocks anyhow, namely width is 1/8th of whole image;
The core of comparison chart picture, extracts grey level histogram.
In described step one, the detailed deterministic process of similar image is:
The first step: the difference of the upper left corner four/a part of mid line region of calculation template image and source images two width image, if difference is close to the threshold value allowed, continues second step, if difference is comparatively large, returns;
Second step: the image difference calculating horizontal line black module, if close, continue the 3rd step, otherwise returns;
3rd step: the image difference calculating vertical line black module, if close, is set to similar classification by this figure, if it is comparatively large to only have first two steps to calculate similar 3rd step gap, is classified as an other class.
Described image retrieval procedure is specially:
After obtaining view data, dynamic cataloging forms multiple particle populations;
When particle populations non-single time, after determining multiple Images Classification, Rapid matching in Images Classification, finds similar image group, returns results and terminates;
When particle populations is single, direct matching image, returns results and terminates.
Described images match process specifically comprises:
When data volume is excessive, when namely data length is greater than threshold value, is divided into two sections of sequences by half, quicksort is carried out to each sequence of differences, select the optimal value come above, terminate and return results;
When data volume is less, when namely data length is less than threshold value, directly carry out quicksort to each sequence of differences, select the optimal value come above, terminate and return results.
A kind of image search method of the present invention, has the following advantages:
A kind of image search method of this invention adopts the mode of new subregion, is no longer that all gray-scale value of acquisition two width image blindly compares, but obtains clockwise according to subregion; Conventional particle colony optimization algorithm is improved, utilizes the most similar image of its function of search selection fast to classify; Classification inside proceeds subregion coupling, and finds out similar image group in order; Improving fast finding algorithm, is on the basis of quicksort thought, with the addition of dynamic logic dividing function, makes quicksort become the fast finding algorithm of intelligence; Retrieval rate is fast, and the imaging accuracy be retrieved is high; Practicality is comparatively strong, applied widely, is easy to promote.
Accompanying drawing explanation
Accompanying drawing 1 is that subregion of the present invention realizes figure.
Accompanying drawing 2 is Modified particle swarm optimization algorithm flow chart of the present invention.
Accompanying drawing 3 is image retrieval process flow diagram of the present invention.
Accompanying drawing 4 is quicksort process flow diagram of the present invention.
Accompanying drawing 5 is practical application figure of the present invention.
Embodiment
Below in conjunction with the drawings and specific embodiments, the invention will be further described.
The invention provides a kind of image search method, in image data base, retrieve similar or identical image according to characteristics of image.First, new subregion thought is utilized first to find similar image sets in source image data storehouse, the classification of similar image is determined by particle group optimizing mode of learning, secondly, in the classification selected, carry out the coupling based on subregion thought by fast finding algorithm, the result of coupling is returned.As shown in accompanying drawing 1 to Fig. 4, its specific implementation process is:
One, template image and source images are extracted grey level histogram information according to partition method, described partition method is specially:
Carrying out subregion to two width images gets a little, becomes the core of both comparisons after some images by every width Iamge Segmentation;
The difference threshold of two width image co-located region is set, then the difference between co-located region corresponding to two images is calculated clockwise, whether the size according to difference completes classification close to threshold value, is categorized as similar image close to threshold value, otherwise is classified as non-similar image;
Two, by find template image the most similar problem change into by improve particle group optimizing classify, the particle group optimizing process of this improvement is:
The present invention improves on the basis of traditional particle cluster algorithm, the i.e. dynamic initialization of population, by the image difference of the given number of random selecting as population, in the image data base of unknown number, retrieve after dynamic initialization, the image collection of convergence is chosen as the first kind after learning certain number of times, then continue studying n time, choose the image collection of convergence as Equations of The Second Kind, with this recurrence until reach amount of images threshold value in classification, in optimizing process, particle is followed current location and velocity optimal particle in colony and is moved, and through obtaining optimum solution by for after iterative search.In each generation, the optimum solution pbest that tracking itself found so far of particle and the optimum solution gbest that finds so far searches for.Formula as follows (1) (2) are the bases that population upgrades:
C1, c2 generally get 2,
get the random number between (0,1), make algorithm have more randomness like this.In particle swarm optimization algorithm, this patent with the addition of random particle and random weights to ensure the verisimilitude of population, the Particle Swarm Optimization Model that different parameters is corresponding different, choosing with convergence for algorithm parameter population is the key factor affecting algorithm performance and efficiency, and have and contact closely, directly affect search procedure and the convergence property of population.
In formula (1), the Section 1 on the formula left side is cognitive item, and the v [] on the right of formula is momentum term, it has been generally acknowledged that particle recognizes experience, thus strengthens study, and remaining item is social item, represents between particle and influences each other, cooperate with each other.Cognitive item represents the competition between particle, and social item represents the cooperation between particle, thus realizes swarm intelligence search;
Three, repeat step one, two, coupling draws the most similar image;
Four, image retrieval is completed by fast finding algorithm, the particular content of this fast finding algorithm is: by analyzing the difference of the histogram information of the partition sections of image and the template image retrieved, to the quicksort that difference portion carries out from small to large, former width images that value is minimum, carry out quicksort, when data length is greater than threshold value time, data being divided equally is two sections, quicksort is carried out, successively recurrence in each section; Finally return and come the most front image, be the image searched.
In described step one, the detailed process of Iamge Segmentation is:
Get the module that two, the centre of image is wide, namely width is 1/16th of whole image;
The mid line region that the centre of getting every a 1/4th images interlocks anyhow, namely width is 1/8th of whole image;
The core of comparison chart picture, extracts grey level histogram.
In described step one, the detailed deterministic process of similar image is:
The first step: the difference of the upper left corner four/a part of mid line region of calculation template image and source images two width image, if difference is close to the threshold value allowed, continues second step, if difference is comparatively large, returns;
Second step: the image difference calculating horizontal line black module, if close, continue the 3rd step, otherwise returns;
3rd step: the image difference calculating vertical line black module, if close, is set to similar classification by this figure, if it is comparatively large to only have first two steps to calculate similar 3rd step gap, is classified as an other class.
Described image retrieval procedure is specially:
After obtaining view data, dynamic cataloging forms multiple particle populations;
When particle populations non-single time, after determining multiple Images Classification, Rapid matching in Images Classification, finds similar image group, returns results and terminates;
When particle populations is single, direct matching image, returns results and terminates.
Described images match process specifically comprises:
When data volume is excessive, when namely data length is greater than threshold value, is divided into two sections of sequences by half, quicksort is carried out to each sequence of differences, select the optimal value come above, terminate and return results;
When data volume is less, when namely data length is less than threshold value, directly carry out quicksort to each sequence of differences, select the optimal value come above, terminate and return results;
In practical application as shown in Figure 5, can complete the retrieving of picture fast, experimentally result judges, the technology of improvement is retrieved in the database that image data amount is larger, and accuracy rate and speed can improve 20%-30%.
Above-mentioned embodiment is only concrete case of the present invention; scope of patent protection of the present invention includes but not limited to above-mentioned embodiment; claims of any a kind of image search method according to the invention and the those of ordinary skill of any described technical field to its suitable change done or replacement, all should fall into scope of patent protection of the present invention.
Claims (5)
1. an image search method, is characterized in that, its specific implementation process is:
One, template image and source images are extracted grey level histogram information according to partition method, described partition method is specially:
Carrying out subregion to two width images gets a little, becomes the core of both comparisons after some images by every width Iamge Segmentation;
The difference threshold of two width image co-located region is set, then the difference between co-located region corresponding to two images is calculated clockwise, whether the size according to difference completes classification close to threshold value, is categorized as similar image close to threshold value, otherwise is classified as non-similar image;
Two, by find template image the most similar problem change into by improve particle group optimizing classify, the particle group optimizing process of this improvement is:
First the dynamic initialization of population is carried out, by the image difference of the given number of random selecting as population;
In image data base, retrieve after dynamic initialization, choose the image collection of convergence after learning certain number of times as the first kind;
Continue studying n time, the image collection choosing convergence is as Equations of The Second Kind, and with this recurrence until reach amount of images threshold value in classification, in optimizing process, particle is followed current location and velocity optimal particle in colony and moved, and through obtaining optimum solution by for after iterative search, finally return classification value;
Three, repeat step one, two, coupling draws the most similar image;
Four, image retrieval is completed by fast finding algorithm, the particular content of this fast finding algorithm is: by analyzing the difference of the histogram information of the partition sections of image and the template image retrieved, to the quicksort that difference portion carries out from small to large, former width images that value is minimum, carry out quicksort, when data length is greater than threshold value time, data being divided equally is two sections, quicksort is carried out, successively recurrence in each section; Finally return and come the most front image, be the image searched.
2. a kind of image search method according to claim 1, is characterized in that, in described step one, the detailed process of Iamge Segmentation is:
Get the module that two, the centre of image is wide, namely width is 1/16th of whole image;
The mid line region that the centre of getting every a 1/4th images interlocks anyhow, namely width is 1/8th of whole image;
The core of comparison chart picture, extracts grey level histogram.
3. a kind of image search method according to claim 2, is characterized in that, in described step one, the detailed deterministic process of similar image is:
The first step: the difference of the upper left corner four/a part of mid line region of calculation template image and source images two width image, if difference is close to the threshold value allowed, continues second step, if difference is comparatively large, returns;
Second step: the image difference calculating horizontal line black module, if close, continue the 3rd step, otherwise returns;
3rd step: the image difference calculating vertical line black module, if close, is set to similar classification by this figure, if it is comparatively large to only have first two steps to calculate similar 3rd step gap, is classified as an other class.
4. a kind of image search method according to claim 1, is characterized in that, described image retrieval procedure is specially:
After obtaining view data, dynamic cataloging forms multiple particle populations;
When particle populations non-single time, after determining multiple Images Classification, Rapid matching in Images Classification, finds similar image group, returns results and terminates;
When particle populations is single, direct matching image, returns results and terminates.
5. a kind of image search method according to claim 4, is characterized in that, described images match process specifically comprises:
When data volume is excessive, when namely data length is greater than threshold value, is divided into two sections of sequences by half, quicksort is carried out to each sequence of differences, select the optimal value come above, terminate and return results;
When data volume is less, when namely data length is less than threshold value, directly carry out quicksort to each sequence of differences, select the optimal value come above, terminate and return results.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410752515.7A CN104361135A (en) | 2014-12-11 | 2014-12-11 | Image retrieval method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410752515.7A CN104361135A (en) | 2014-12-11 | 2014-12-11 | Image retrieval method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104361135A true CN104361135A (en) | 2015-02-18 |
Family
ID=52528395
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410752515.7A Pending CN104361135A (en) | 2014-12-11 | 2014-12-11 | Image retrieval method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104361135A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105095483A (en) * | 2015-08-14 | 2015-11-25 | 北京铭嘉实咨询有限公司 | Image code recognition method and system |
| CN105469120A (en) * | 2015-12-15 | 2016-04-06 | 浪潮电子信息产业股份有限公司 | Image matching method and apparatus |
| CN107145510A (en) * | 2017-03-31 | 2017-09-08 | 西安科技大学 | A kind of mathematical formulae searching method and device |
| CN107657037A (en) * | 2017-09-29 | 2018-02-02 | 郑州云海信息技术有限公司 | A kind of efficient image search method |
| CN108764258A (en) * | 2018-05-24 | 2018-11-06 | 西安电子科技大学 | A kind of optimum image collection choosing method being inserted into for group's image |
| CN114027854A (en) * | 2021-12-01 | 2022-02-11 | 杭州电子科技大学 | BECT spike detection method based on optimal template matching and morphological feature extraction |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7599894B2 (en) * | 2005-03-04 | 2009-10-06 | Hrl Laboratories, Llc | Object recognition using a cognitive swarm vision framework with attention mechanisms |
| CN102842043A (en) * | 2012-07-17 | 2012-12-26 | 西安电子科技大学 | Particle swarm classifying method based on automatic clustering |
| CN102902826A (en) * | 2012-11-08 | 2013-01-30 | 公安部第三研究所 | Quick image retrieval method based on reference image indexes |
| US8645294B1 (en) * | 2004-02-03 | 2014-02-04 | Hrl Laboratories, Llc | Method for image registration utilizing particle swarm optimization |
| CN103824086A (en) * | 2014-03-24 | 2014-05-28 | 东方网力科技股份有限公司 | Image matching method and device |
-
2014
- 2014-12-11 CN CN201410752515.7A patent/CN104361135A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8645294B1 (en) * | 2004-02-03 | 2014-02-04 | Hrl Laboratories, Llc | Method for image registration utilizing particle swarm optimization |
| US7599894B2 (en) * | 2005-03-04 | 2009-10-06 | Hrl Laboratories, Llc | Object recognition using a cognitive swarm vision framework with attention mechanisms |
| CN102842043A (en) * | 2012-07-17 | 2012-12-26 | 西安电子科技大学 | Particle swarm classifying method based on automatic clustering |
| CN102902826A (en) * | 2012-11-08 | 2013-01-30 | 公安部第三研究所 | Quick image retrieval method based on reference image indexes |
| CN103824086A (en) * | 2014-03-24 | 2014-05-28 | 东方网力科技股份有限公司 | Image matching method and device |
Non-Patent Citations (5)
| Title |
|---|
| 周鲜成: ""基于微粒群优化的图像分类方法研究"", 《中文核心期刊《微计算机信息》(管控一体化》 * |
| 易文晟: "《图像语义检索和分类技术研究》", 《中国博士学位论文全文数据库 信息科技辑》 * |
| 杨文涛 等: ""微粒群算法在图像检索中的应用"", 《华北水利水电学院学报》 * |
| 胡浩 等: ""基于颜色、边缘综合特征的图像检索技术研究"", 《计算机工程与设计》 * |
| 赵子鹏: ""粒子群优化算法及其在图像检索中相关反馈上的应用"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105095483A (en) * | 2015-08-14 | 2015-11-25 | 北京铭嘉实咨询有限公司 | Image code recognition method and system |
| CN105469120A (en) * | 2015-12-15 | 2016-04-06 | 浪潮电子信息产业股份有限公司 | Image matching method and apparatus |
| CN107145510A (en) * | 2017-03-31 | 2017-09-08 | 西安科技大学 | A kind of mathematical formulae searching method and device |
| CN107657037A (en) * | 2017-09-29 | 2018-02-02 | 郑州云海信息技术有限公司 | A kind of efficient image search method |
| CN108764258A (en) * | 2018-05-24 | 2018-11-06 | 西安电子科技大学 | A kind of optimum image collection choosing method being inserted into for group's image |
| CN108764258B (en) * | 2018-05-24 | 2022-03-04 | 西安电子科技大学 | Optimal image set selection method for group image insertion |
| CN114027854A (en) * | 2021-12-01 | 2022-02-11 | 杭州电子科技大学 | BECT spike detection method based on optimal template matching and morphological feature extraction |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113516012B (en) | Pedestrian re-identification method and system based on multi-level feature fusion | |
| CN108647577B (en) | Self-adaptive pedestrian re-identification method and system for difficult excavation | |
| CN105844669B (en) | A kind of video object method for real time tracking based on local Hash feature | |
| CN104361135A (en) | Image retrieval method | |
| CN109389180A (en) | A power equipment image-recognizing method and inspection robot based on deep learning | |
| CN103617217A (en) | Hierarchical index based image retrieval method and system | |
| CN110188225B (en) | Image retrieval method based on sequencing learning and multivariate loss | |
| CN112085072B (en) | Cross-modal retrieval method of sketch retrieval three-dimensional model based on space-time characteristic information | |
| CN102254015A (en) | Image retrieval method based on visual phrases | |
| CN107316053A (en) | A kind of cloth image Rapid matching search method | |
| CN111325264A (en) | Multi-label data classification method based on entropy | |
| CN105046714A (en) | Unsupervised image segmentation method based on super pixels and target discovering mechanism | |
| Liu et al. | Analysis of anchor-based and anchor-free object detection methods based on deep learning | |
| CN102708367A (en) | Image identification method based on target contour features | |
| CN109871379A (en) | A kind of online Hash K-NN search method based on data block study | |
| Guo | Research on sports video retrieval algorithm based on semantic feature extraction | |
| CN109472712A (en) | A kind of efficient Markov random field Combo discovering method strengthened based on structure feature | |
| CN109242039A (en) | It is a kind of based on candidates estimation Unlabeled data utilize method | |
| CN117763185A (en) | Hash image retrieval method based on thinking space dimension | |
| CN110334226B (en) | Depth image retrieval method fusing feature distribution entropy | |
| CN114168782B (en) | Deep hash image retrieval method based on triplet network | |
| CN110532867A (en) | A kind of facial image clustering method based on Fibonacci method | |
| CN116011564A (en) | Entity relationship completion method, system and application for power equipment | |
| Li et al. | Region proposal ranking via fusion feature for object detection | |
| CN108090514B (en) | Infrared image identification method based on two-stage density clustering |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150218 |