CN103116754B - Batch images dividing method and system based on model of cognition - Google Patents
Batch images dividing method and system based on model of cognition Download PDFInfo
- Publication number
- CN103116754B CN103116754B CN201310029906.1A CN201310029906A CN103116754B CN 103116754 B CN103116754 B CN 103116754B CN 201310029906 A CN201310029906 A CN 201310029906A CN 103116754 B CN103116754 B CN 103116754B
- Authority
- CN
- China
- Prior art keywords
- image
- model
- user
- cognition
- prospect
- 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.)
- Active
Links
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides a kind of batch images dividing method and system based on model of cognition. The function that application on current many desktop picture process softwares or even some mobile devices all provides image to cut apart, for separating the prospect of a width picture with background. But this class software needs user manually to iris out the profile of prospect in picture conventionally, and user's interactive operation is had relatively high expectations, and cannot carry out batch process to multiple images. The present invention is directed to the problems referred to above design, for user provides the multiple different model based on identification, user only needs manually to select one or more models, and then chooses picture to be processed, and system can be carried out cutting operation to all pictures automatically according to user-selected model.
Description
Technical field
The present invention relates to image processing techniques and computer vision field, relate in particular to a kind of from picture based on selecteed knowledgeMethod and the system thereof of the part of other model auto Segmentation and Model Matching.
Background technology
Image is cut apart the process that image is divided into multiple parts according to certain rule or object that refers to, for example, by a width scenery with hills and watersThe image of drawing is divided into the part that comprises mountain, the part that comprises water and other parts, and for example by the people in the image of a width portraitThing is with background segment process out. Image Segmentation Technology is widely used in practice, and application comprises medical image, satelliteLocation and the landscaping treatment of image later stage etc.
In recent years, along with the develop rapidly of software and hardware technology, the image processing software on some personal computers or even someMobile phone is provided by the function that all provides image to cut apart, can person easy to use to the photo of taking carry out post-processed (remove foreign material,Image Mosaics etc.). But present stage this part function sight line all need user manual in image, iris out the portion that need to cut apartPoint profile, and the order of accuarcy of the profile irised out of the result cut apart of image and user has very large relation, this is to useThe interactive operation at family has proposed very high requirement. In addition, need user manually on image the operation of delineate also determineExisting image processing software cannot carry out image dividing processing in batches to multiple images. Therefore, in the urgent need to developing onePlant image and cut apart module, it has the function that only needs user's shirtsleeve operation can carry out to multiple images batch splitting processing.
Batch images based on model of cognition is cut apart, as its name suggests, be exactly user choose want the model of cognition that is partitioned into andNeed to carry out, after one or more picture of image cutting operation, adopting the algorithm of a series of machine learning and pattern-recognition, to dividingIn the picture cutting, mate part to be matched and judge, and the prospect after cutting apart and background data are presented in software to userCan further operate the image after cutting apart, as deletion or preservation etc. This mode has greatly been simplified user's behaviourDo, and saved the valuable time for user.
Summary of the invention
Cannot carry out to picture the shortcoming of batch images dividing processing in order to overcome conventional images process software, provide a kind of based onModel of cognition, the corresponding position of batch picture model is carried out Auto-matching and cut apart, and by the prospect being partitioned into and backgroundData are presented in software, facilitate user to carry out the method and system of post-processed.
The invention provides a kind of method that batch images based on model of cognition is cut apart, comprise the following steps:
1) system enters Model Selection pattern;
2) system obtains user-selected model of cognition, enters picture preference pattern;
3) system obtains user-selected picture;
4) system is based on 2) model of cognition that obtains in step, to 3) the image data operation batch images that obtains in step dividesCut algorithm, obtain the part being partitioned in picture;
5) result is sent to output unit by system;
The invention provides a kind of batch images partitioning algorithm based on model of cognition, this algorithm is based on user-selected oneOr multiple locally variable models and corresponding characteristics of image, to the different details under different resolution in one or more picture respectivelyMate and mark, finally appraisal result being added.
The invention provides the system that a kind of batch images based on model of cognition is cut apart, comprise with lower unit:
1) receiving element, receives model of cognition and picture that user selects;
2) image cutting unit, carries out the image cutting procedure of core;
3) output unit, exports the result after cutting apart.
Advantage of the present invention is:
1) having overcome conventional images process software cannot carry out the shortcoming that image is cut apart to batch picture, has filled up current this neckThe technological gap in territory;
2) further promoted the function of image processing software, provided convenience for the processing of user's comparison film, saved after userThe time that phase processes;
3) adopted the algorithm of a series of machine learning and pattern-recognition, in comparison film, matching area is selected and judges, largeAmplitude has improved automaticity and the efficiency of system.
Brief description of the drawings
Fig. 1 is the workflow diagram of a kind of batch images dividing method based on model of cognition provided by the invention;
Fig. 2 is the workflow diagram of a kind of batch images partitioning algorithm based on model of cognition provided by the invention.
Fig. 3 is the structural representation of a kind of batch images segmenting system based on model of cognition provided by the invention;
Detailed description of the invention
Below in conjunction with the accompanying drawing in the present invention, technical scheme of the present invention is carried out to clear, intactly description. Based on thisEmbodiment in bright, the every other enforcement that those of ordinary skill in the art obtain under the prerequisite of not doing creative workExample, all belongs to the scope of protection of the invention.
For making object of the present invention, technical scheme and advantage clearer, next with reference to the accompanying drawings to the invention processExample is described in detail.
As shown in Figure 1, the workflow of the batch images dividing method based on model of cognition comprises following step:
User selects the model of cognition that will mate on the control panel of image processing software, and system obtains user-selectedModule, then enter picture select interface;
User selects interface to select to carry out the picture of image dividing processing at picture;
System obtains raw image data and the model of cognition data that input block transmits;
System is the model of cognition operation image partitioning algorithm based on user-selected to the raw image data of every pictures successively,The region of mating with model in detected image, its operation principle is as shown in Figure 2;
System is sent to output unit to image dividing processing result.
As shown in Figure 2, the batch images partitioning algorithm based on model of cognition is divided into following step:
Step 1, obtains raw image data, and the training of process that user selects is stored in the multistage part in storage mediumCan varying model;
Step 2, extracts the feature in image, and what in the present invention, adopt is a kind of GIST feature of 15X1536 dimension;
Step 3, is used different model of cognition, mates with the feature of extracting respectively, and to matching result weighted sumAcquisition prospect mask image (mask);
Step 4, according to mask image initial foreground/background model. ;
Step 5, prospect of the application/background model is carried out binary segmentation to image, and segmentation result and foreground/background model are carried outRelatively, if convergence is final result, otherwise, iterative step 5;
Step 6, exports final segmentation result.
Shown in Fig. 3, a kind of batch images segmenting system based on model of cognition of the invention process, comprises with lower unit:
Receiving element, model of cognition data and the image data (being original image information) selected for receiving user, and willModel data and view data transfer to image cutting unit to process.
Image cutting unit, for detection of the part of mating with selected module in image, result (be partitioned into beforeScape and background data) transfer to output unit.
Output unit, output detections result is in software.
Finally, it should be pointed out that above embodiment is only the more representational example of the present invention. Obviously, technical side of the present inventionCase is not limited to above-described embodiment, can also have many distortion. Those of ordinary skill in the art can not depart from of the present inventionUnder the bright state of mind, make various modifications or variation for above-described embodiment, thereby protection scope of the present invention is not by above-mentioned realityExecute example institute and limit, and should be the maximum magnitude that meets the inventive features that claims mention.
Claims (5)
1. the image partition method based on model of cognition, is characterized in that, comprises the following steps:
1) provide multiple model of cognition to select for user;
2) obtain the model of cognition that user manually selects;
3) enter picture and choose pattern, obtain the image processed of need that user selects;
4) model based on user-selected, the image that user is selected moves the image segmentation algorithm based on model of cognition, obtainsThe part matching with each model in image, is divided into prospect and background by image; Comprise the following steps:
(41) by collecting the picture on a large amount of internets, form picture database, by owning in training picture databasePicture forms different model of cognition, selects for user, and user, according to selected image, selects suitable model of cognition;
(42) use different model of cognition respectively image to be detected, identify the foreground area of each class model in image,And use the position in the each region of detection window mark;
(43) image-region to each detection window institute mark, calculates the characteristics of image in this region;
(44) characteristics of image of collection All Ranges, mates one by one with training characteristics, and the training characteristics is here deposited before beingStorage is in system, by training image being detected and obtaining collection, the knot of feature after coupling by feature extractionFruit is weighted addition, obtains Potential Prediction mask, uses 1 or 0 mark prospect and background area;
(45) according to the prediction mask of prospect, set up the gauss hybrid models of prospect and the background of image, and according to the mould of setting upThe prospect of type estimated image, if estimated result converges to certain piece region, this region is final image segmentation result;
5) prospect after cutting apart and background data are sent to output unit.
2. right to use requires a system for the method described in 1, it is characterized in that, comprises following three unit that connect successively:
1) input block, receives the picture processed of need and corresponding model of cognition that user selects, and is sent to image and cuts apartUnit;
2) image cutting unit, transmits image and the model of cognition of coming according to input block, first identify detected imageIn the corresponding region of user-selected model; Then carry out image cutting operation, the prospect after cutting apart and background result are sent toOutput unit;
3) output unit, exports the result after cutting apart, and user can carry out after other the prospect after cutting apart and background imagePhase operation.
3. system as claimed in claim 2, is characterized in that input block 1), comprise memory module and user interactive module, with figureBe connected as cutting unit, the model data and the view data that receive user's selection transfer to image cutting unit to process.
4. system as claimed in claim 2, is characterized in that image cutting unit 2), comprise detection module, Model Identification coupling mouldPiece and image are cut apart module, and with input block, output unit is connected, and the data that reception input block imports into are also cut apart placeAfter reason, the prospect after cutting apart and background data are transferred to output unit.
5. system as claimed in claim 2, is characterized in that output unit 3), comprise display module and memory module, segmentation resultCan be used as image and show, storage simultaneously is further carried out other operations for user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310029906.1A CN103116754B (en) | 2013-01-24 | 2013-01-24 | Batch images dividing method and system based on model of cognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310029906.1A CN103116754B (en) | 2013-01-24 | 2013-01-24 | Batch images dividing method and system based on model of cognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103116754A CN103116754A (en) | 2013-05-22 |
| CN103116754B true CN103116754B (en) | 2016-05-18 |
Family
ID=48415126
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310029906.1A Active CN103116754B (en) | 2013-01-24 | 2013-01-24 | Batch images dividing method and system based on model of cognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103116754B (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9679224B2 (en) * | 2013-06-28 | 2017-06-13 | Cognex Corporation | Semi-supervised method for training multiple pattern recognition and registration tool models |
| CN104346506B (en) * | 2013-07-31 | 2018-11-30 | Ge医疗***环球技术有限公司 | The method and device that a kind of pair of medical image image is handled and exported |
| CN103440738B (en) * | 2013-08-16 | 2016-06-29 | 吉林大学 | A kind of overcrowding method for early warning of tourist footbridge in scenic area |
| CN107305676B (en) * | 2016-04-20 | 2021-07-13 | 上海恒名软件有限公司 | Method for realizing standard comparison of store display commodities in batch shopping software |
| CN107305675B (en) * | 2016-04-20 | 2021-07-13 | 上海恒名软件有限公司 | Shopping software display ordering method |
| CN107025457B (en) * | 2017-03-29 | 2022-03-08 | 腾讯科技（深圳）有限公司 | Image processing method and device |
| CN107480587B (en) * | 2017-07-06 | 2020-12-08 | 创新先进技术有限公司 | Model configuration and image recognition method and device |
| CN107516090B (en) | 2017-09-11 | 2021-09-17 | 北京百度网讯科技有限公司 | Integrated face recognition method and system |
| CN107633219A (en) * | 2017-09-11 | 2018-01-26 | 北京百度网讯科技有限公司 | Integrated optical character identifying method and system |
| CN109377509B (en) * | 2018-09-26 | 2021-03-26 | 达闼机器人有限公司 | Image semantic segmentation labeling method and device, storage medium and equipment |
| CN110004664B (en) * | 2019-04-28 | 2021-07-16 | 深圳数联天下智能科技有限公司 | Clothes stain recognition method and device, washing machine and storage medium |
| US11024027B2 (en) * | 2019-09-13 | 2021-06-01 | Siemens Healthcare Gmbh | Manipulable object synthesis in 3D medical images with structured image decomposition |
| CN112008982B (en) * | 2020-08-14 | 2023-03-21 | 珠海赛纳三维科技有限公司 | Model printing device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101558404A (en) * | 2005-06-17 | 2009-10-14 | 微软公司 | Image segmentation |
| CN101655909A (en) * | 2008-08-21 | 2010-02-24 | 索尼（中国）有限公司 | Device and method for calculating matching degree |
| CN102054270A (en) * | 2009-11-10 | 2011-05-11 | 华为技术有限公司 | Method and device for extracting foreground from video image |
-
2013
- 2013-01-24 CN CN201310029906.1A patent/CN103116754B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101558404A (en) * | 2005-06-17 | 2009-10-14 | 微软公司 | Image segmentation |
| CN101655909A (en) * | 2008-08-21 | 2010-02-24 | 索尼（中国）有限公司 | Device and method for calculating matching degree |
| CN102054270A (en) * | 2009-11-10 | 2011-05-11 | 华为技术有限公司 | Method and device for extracting foreground from video image |
Non-Patent Citations (2)
| Title |
|---|
| Photoshop抠图技术分析与应用;周贤善;《现代计算机》;20110531;全文 * |
| 图像前景提取技术研究;高正伟;《中国优秀硕士学位论文全文数据库》;20080715;第47-61页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103116754A (en) | 2013-05-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103116754B (en) | Batch images dividing method and system based on model of cognition | |
| US11430205B2 (en) | Method and apparatus for detecting salient object in image | |
| CN107862698B (en) | Light field foreground segmentation method and device based on K mean cluster | |
| CN109816663B (en) | Image processing method, device and equipment | |
| CN109961453B (en) | Image processing method, device and equipment | |
| CN110163076B (en) | Image data processing method and related device | |
| US9665962B2 (en) | Image distractor detection and processng | |
| US20150187076A1 (en) | System and Methods for Persona Identification Using Combined Probability Maps | |
| KR102190527B1 (en) | Apparatus and method for automatic synthesizing images | |
| CN103455984A (en) | Method and device for acquiring Kinect depth image | |
| CN109635783A (en) | Video monitoring method, device, terminal and medium | |
| CN108388882A (en) | Based on the gesture identification method that the overall situation-part is multi-modal RGB-D | |
| CN111127307A (en) | Image processing method, image processing device, electronic equipment and computer readable storage medium | |
| CN111080670A (en) | Image extraction method, device, equipment and storage medium | |
| CN104951440B (en) | Image processing method and electronic equipment | |
| KR20180067909A (en) | Apparatus and method for segmenting image | |
| CN113409342A (en) | Training method and device for image style migration model and electronic equipment | |
| Chen et al. | Visual depth guided image rain streaks removal via sparse coding | |
| CN104408444A (en) | Human body action recognition method and device | |
| CN106815237B (en) | Search method, search device, user terminal and search server | |
| CN104318251B (en) | A kind of photograph real-time processing method and device | |
| CN111383289A (en) | Image processing method, image processing device, terminal equipment and computer readable storage medium | |
| CN113409331B (en) | Image processing method, image processing device, terminal and readable storage medium | |
| WO2023283894A1 (en) | Image processing method and device | |
| CN104506768A (en) | Method and device for image selection as well as terminal |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |