CN103577475A - Picture automatic sorting method, picture processing method and devices thereof - Google Patents
Picture automatic sorting method, picture processing method and devices thereof Download PDFInfo
- Publication number
- CN103577475A CN103577475A CN201210276320.0A CN201210276320A CN103577475A CN 103577475 A CN103577475 A CN 103577475A CN 201210276320 A CN201210276320 A CN 201210276320A CN 103577475 A CN103577475 A CN 103577475A
- Authority
- CN
- China
- Prior art keywords
- picture
- classification
- sorted
- feature
- characteristic
- 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
Images
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
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/5846—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using extracted text
-
- 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
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/5838—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using colour
Abstract
The embodiment of the invention discloses a picture automatic sorting method comprising the following steps of receiving pictures to be sorted; reading feature categories in a feature library; extracting feature data of the pictures to be sorted according to the feature categories; matching the extracted feature data with corresponding preset feature data of the feature categories, and sorting the pictures to be sorted, of which the feature data can be matched with, into a category. The embodiment of the invention further provides a picture processing method based on the picture sorting, a picture automatic sorting device and a picture processing device based on the picture sorting. The embodiment of the invention improves the picture sorting efficiency and accuracy.
Description
Technical field
The application relates to picture Processing Technique field, particularly a kind of mass picture automatic partition class methods, the image processing method based on picture classification, and each self-corresponding device.
Background technology
Along with the development of Internet technology, information magnanimity increases, and network picture and image data be rapid growth also.Picture, as a kind of important information source (carrier), need to carry out centralized stores, processing and some personalized application conventionally.But that these pictures are often originated is different, form is different, picture form thousand is poor outer.For meeting the normal utilization to picture, usually need these mass pictures to classify, to carry out differential management (processing).
Picture classification mode of the prior art is to adopt artificial cognition mechanism, by hand labor, a large amount of pictures is carried out to class discrimination and merger one by one according to preset rules.Although this mode is succinct, convenient,, need to waste a large amount of human and material resources and financial cost, classification effectiveness is lower, during especially in the face of mass picture, all the more so.And artificial cognition mode is owing to constantly repeating machinery work, easily tired, causes classifying inaccurate, thereby cannot meet practical application needs separately.
Summary of the invention
Because the problem that prior art exists, the embodiment of the present application provides a kind of picture automatic partition class methods, the image processing method based on picture classification and has installed accordingly, to realize picture mechanized classification and processing, improve efficiency and the accuracy of picture classification and processing.
The picture automatic partition class methods that the embodiment of the present application provides comprise:
Receive picture to be sorted;
Read the feature classification in feature database;
According to described feature classification, extract the characteristic of described picture to be sorted;
The characteristic default characteristic corresponding with described feature classification that coupling is extracted, the picture merger to be sorted that characteristic can be mated is a class.
Preferably, described feature classification comprises picture background, picture prospect, picture character and/or picture personage.
Further preferably, described feature classification is picture background and/or picture prospect, the described default characteristic corresponding with feature classification comprises background color number, the background colour of picture background, and/or, the foreground color number of picture prospect, foreground area number, the prospect band of position: the described characteristic according to the described picture to be sorted of feature classification extraction comprises:
Described picture to be sorted is divided into one or more subgraphs;
Extract background color number, the background colour of each subgraph, and/or, foreground color number, foreground area number, the foreground area position of extracting each subgraph.
Further preferably, foreground color number, foreground area number, the foreground area position of extraction subgraph specifically comprise:
The pixel number of every kind of color in the default color set of statistics within the scope of described subgraph;
Calculate according to the following formula the significance value of every kind of color:
In formula: c
ifor i element in default color set, the element number that N is color set, w
ifor the pixel number of the i kind color in default color set within the scope of subgraph, dist (c
kc
i) be the distance between two kinds of colors in color set;
In subgraph, the color significance value of each pixel is as the significance value of this pixel;
The region that pixel significance value sum is greater than to predetermined threshold value is identified as foreground area;
According to the foreground area of identification, determine number, the foreground area position of foreground color number, foreground area.
Preferably, described feature classification is picture character, the described default characteristic corresponding with feature classification comprises character area position, character area size and/or word content: the described characteristic according to the described picture to be sorted of feature classification extraction comprises:
Described picture to be sorted is carried out to character area location, to determine character area position and/or character area size; And/or,
Described picture to be sorted is carried out to character area and locate laggard style of writing word identification, to determine the word content of character area.
Further preferably, described picture to be sorted is carried out to character area location and specifically comprise: by carrying out character area location based on texture statistics method or based on region analysis method, and/or,
Described picture to be sorted is carried out to word identification specifically to be comprised: by KD, set method of identification the word in character area is identified.
Preferably, described feature classification is picture personage, the described default characteristic corresponding with feature classification comprises people's face number, people's face position and/or face complexion: described people's face number, people's face position and/or the face complexion that extracts described picture to be sorted that be characterized as that extracts described picture to be sorted according to feature classification.
The embodiment of the present application also provides a kind of image processing method based on picture classification.The method comprises:
Receive picture to be sorted;
Read the feature classification in feature database;
According to described feature classification, extract the characteristic of described picture to be sorted;
The characteristic default characteristic corresponding with described feature classification that coupling is extracted, the picture merger to be sorted that characteristic can be mated is a class;
The picture of a classification is processed according to default processing rule.
Preferably, the feature classification in described default processing rule and feature database has corresponding relation, described the picture of a classification is processed and is specially according to default processing rule:
The picture of a classification is processed according to default processing rule corresponding to the feature classification with reading from feature database.
Preferably, described the picture of a classification processed and is specially according to default processing rule:
Predeterminated position at the picture of a classification adds presupposed information.
The embodiment of the present application also provides a kind of picture mechanized classification device.This device comprises: receiving element, reading unit, extraction unit, matching unit and classification unit, wherein:
Described receiving element, for receiving picture to be sorted;
Described reading unit, for reading the feature classification in feature database;
Described extraction unit, for extracting the characteristic of described picture to be sorted according to the described feature classification reading;
Described matching unit, for mating the picture feature data default characteristic corresponding with described feature classification of extraction;
Described classification unit is a class for the picture merger to be sorted that characteristic can be mated.
Preferably, described feature classification is picture background and/or picture prospect, described and feature classification characteristic of correspondence data comprise background color number, the background colour of picture background, and/or, the foreground color number of picture prospect, foreground area number, foreground area position, described extraction unit comprises: divide subelement and extract subelement, wherein:
Described division unit, for being divided into one or more subgraphs by described picture to be sorted;
Described extraction subelement, for extract background color number, the background colour of each subgraph according to the described feature classification reading, and/or, foreground color number, foreground area number, the foreground area position of extracting each subgraph.
Further preferably, described extraction subelement is when extracting foreground color number, foreground area number, the foreground area position of subgraph, this subelement comprises: pixel statistics subelement, significance value computation subunit, foreground area recognin unit and definite subelement, wherein:
Described pixel statistics subelement, for adding up every kind of color in default color set pixel number within the scope of described subgraph;
Described significance value computation subunit, for calculating according to the following formula the significance value of every kind of color, and in subgraph the color significance value of each pixel as the significance value of this pixel:
In formula: c
ifor i element in default color set, the element number that N is color set, w
ifor the pixel number of the i kind color in default color set within the scope of subgraph, dist (c
kc
i) be the distance between two kinds of colors in color set;
Described foreground area recognin unit, is identified as foreground area for pixel significance value sum being greater than to the region of predetermined threshold value;
Described definite subelement, for determining foreground color number, foreground area number, foreground area position according to the foreground area of identification.
Preferably, described feature classification is picture character, and the described default characteristic corresponding with feature classification comprises character area position, character area size and/or word content, described extraction unit comprises: locator unit, and/or, recognin unit, wherein:
Described locator unit, for described picture to be sorted is carried out to character area location, to determine character area position and/or character area size;
Described recognin unit, locates laggard style of writing word identification for described picture to be sorted being carried out to character area, to determine the word content of character area.
Further preferably, described locator unit is by carrying out character area location based on texture statistics method or based on region analysis method, and/or described recognin unit is set method of identification by KD the word in character area is identified.
The embodiment of the present application also provides a kind of picture processing device based on picture classification.This device comprises: receiving element, reading unit, extraction unit, matching unit, classification unit and processing unit, wherein:
Described receiving element, for receiving picture to be sorted;
Described reading unit, for reading the feature classification in feature database;
Described extraction unit, for extracting the characteristic of described picture to be sorted according to the described feature classification reading;
Described matching unit, for mating the picture feature data default characteristic corresponding with described feature classification of extraction;
Described classification unit is a class for the picture merger to be sorted that characteristic can be mated;
Described processing unit, for processing according to preset rules the picture of a classification.
The embodiment of the present application is after receiving picture to be sorted, according to the feature classification in feature database, extract the characteristic of picture to be sorted, then these characteristics default characteristic corresponding with feature classification being mated, if can mate, is a class by its merger.Compared with prior art, the embodiment of the present application can automatic acquisition feature classification and with feature classification characteristic of correspondence data, and the characteristic of acquisition is mated to distinguish classification with the characteristic of extraction, thereby can automatically complete the classification of mass picture, save human and material resources and financial cost, improved the work efficiency of picture classification.And, owing to taking robotization treatment measures, as long as pre-set feature classification and characteristic, can by one by one coupling realize picture classification, thereby improved the accuracy of picture classification.The picture classification mode of the embodiment of the present application high-level efficiency and pin-point accuracy has met the application demand of various occasions preferably.
Accompanying drawing explanation
In order to be illustrated more clearly in the embodiment of the present application or technical scheme of the prior art, to the accompanying drawing of required use in embodiment or description of the Prior Art be briefly described below, apparently, the accompanying drawing the following describes is only some embodiment that record in the application, for those of ordinary skills, do not paying under the prerequisite of creative work, can also obtain according to these accompanying drawings other accompanying drawing.
Fig. 1 is the picture mechanized classification method flow diagram of the embodiment of the present application one;
Fig. 2 is the picture mechanized classification method flow diagram of the embodiment of the present application two;
Fig. 3 (a)～(f) is the effect schematic diagram of embodiment described in Fig. 2;
Fig. 4 is the structured flowchart of the picture mechanized classification device of the embodiment of the present application three;
Fig. 5 is the structured flowchart of the picture processing device based on picture classification of the embodiment of the present application four.
Embodiment
In order to make those skilled in the art person understand better the technical scheme in the application, below in conjunction with the accompanying drawing in the embodiment of the present application, technical scheme in the embodiment of the present application is clearly and completely described, obviously, described embodiment is only the application's part embodiment, rather than whole embodiment.Embodiment based in the application, those of ordinary skills are not making the every other embodiment obtaining under creative work prerequisite, all should belong to the scope of the application's protection.
For the application's above-mentioned purpose, feature and advantage can be become apparent more, below in conjunction with the drawings and specific embodiments, the application is described in further detail.
Embodiment mono-
Referring to Fig. 1, the figure shows the flow process of the picture automatic classification method of the embodiment of the present application one.The present embodiment comprises:
Step S101: receive picture to be sorted;
Picture to be sorted is that the embodiment of the present application is prepared the figure film source of classifying and processing, this figure film source can be a sheet by a sheet independently picture receiving by network, can be the picture being cached in advance in picture library, the source that the embodiment of the present application is treated category images impose any restrictions yet.While receiving picture to be sorted, can once only receive a pictures, also can once receive plurality of pictures, this depends on processing power and the requirement to picture classification device work efficiency of picture classification device.
Step S102: read the feature classification in feature database;
Feature database is to be exclusively used in to place feature class destination data storehouse.In actual application, feature classification characteristic of correspondence data can also be stored in this feature database in advance, also can in picture classification processing procedure, eject dialog box, require user's input feature vector classification characteristic of correspondence data.User can select wherein any mode according to the needs of oneself, the former is due to pre-set unified characteristic, thereby being conventionally applicable to large batch of picture automatic classification, the latter is owing to can adjusting at any time characteristic, thereby can realize multi-level, the personalized classification of picture.
Step S103: the characteristic of extracting described picture to be sorted according to described feature classification;
Obtain after feature classification, for ease of carrying out the coupling of characteristic, need to from picture to be sorted, extract characteristic of correspondence data.Different feature classifications, the extracting method of its characteristic may there are differences, such as extracting the background data of picture, can adopt picture cutting technique and picture marking area extractive technique, the lteral data that extracts picture can adopt picture character location and character recognition technology.
Step S104: the characteristic default characteristic corresponding with described feature classification that coupling is extracted, the picture merger to be sorted that characteristic can be mated is a class.
By aforementioned characteristic extraction step, extract after the characteristic of picture to be sorted, these data default characteristic corresponding with feature classification can be mated, here " the default characteristic corresponding with feature classification ", as previously mentioned, can be pre-stored in feature database, from feature database, reading when needed, can be also interim setting.When matching operation is specifically carried out, can be to extract disposable coupling after the picture to be sorted of some, can be also matching process of a feature extraction.By coupling, picture to be sorted can be divided into at least two classes: the unmatched picture to be sorted of the picture to be sorted that characteristic matches and characteristic.When feature classification is stagewise feature classification, the classification of division can be multistage.The picture number that is integrated into a class is how much relevant with figure film source and feature classification and corresponding default characteristic, in figure film source fixedly time, if user need to carry out multiscale category division by picture, feature classification can be carried out to refinement, default many levels, a plurality of rank; If user is higher to category images accuracy requirement, can improve the standard of characteristic, or characteristic is set to " section ", rather than single " numerical value ".These different implementations depend on the various real needs of practical application.
The present embodiment is after receiving picture to be sorted, according to the feature classification in feature database, extract the characteristic of picture to be sorted, then these characteristics default characteristic corresponding with feature classification being mated, if can mate, is a class by its merger.Compared with prior art, the present embodiment can automatic acquisition feature classification and with feature classification characteristic of correspondence data, and the characteristic of acquisition is mated to distinguish classification with the characteristic of extraction, thereby can automatically complete the classification of mass picture, save human and material resources and financial cost, improved the work efficiency of picture classification.And, owing to taking robotization treatment measures, as long as pre-set feature classification and characteristic, can by one by one coupling realize picture classification, thereby improved the accuracy of picture classification.The picture classification mode of the embodiment of the present application high-level efficiency and pin-point accuracy has met the application demand of various occasions preferably.In addition, according to the various concrete scene of practical application, can realize multi-level picture classification by many features classification, by characteristic high standard, range selector setting, realize high-precision picture classification.
Embodiment bis-
For further illustrating the application's technical scheme, with a comparatively concrete example, describe below.In this embodiment, feature classification comprises: picture background, picture prospect, picture character and picture personage, the default characteristic that picture background feature classification is corresponding is background color number, background colour, the characteristic of picture prospect is foreground object number, foreground object position, picture character characteristic of correspondence data are character area position, character area size and word content, picture personage characteristic of correspondence data people face number, people's face position and face complexion.The value of these feature classifications and feature classification characteristic of correspondence data thereof is pre-stored among feature database.
Referring to Fig. 2, the figure shows the flow process of the picture automatic partition class methods of the present embodiment two.This embodiment comprises:
Step S201: read picture background, picture prospect, picture character and four feature classifications of picture personage and characteristic of correspondence data thereof in feature database;
It is because the present embodiment adopts pre-stored characteristic in feature database that the present embodiment is placed on first step execution by the operation of reading feature classification and characteristic, characteristic need to be set in picture classification processing procedure as required temporarily.
Step S202: receive a picture to be sorted;
Step S203: two background characteristics of background complexity, background colour of extracting described picture to be sorted, extract three foreground features data in prospect complexity, foreground area number, foreground area position of described picture to be sorted, extract character area position, character area size and three character features data of word content of picture to be sorted, and people's face number, people's face position and three character features data of face complexion of extracting picture to be sorted;
As previously mentioned, different according to feature classification, the method for extracting this feature class object characteristic of picture to be sorted may be different.Such as, for the characteristic of extracting picture prospect and background characteristics project, can first adopt picture segmentation technology that picture is divided into a plurality of subgraphs, then take picture marking area extractive technique to extract the color number of colours of picture and foreground area number, foreground area position, on this technical foundation, also can extract picture background look etc.; For extracting picture character characteristic item object characteristic, can adopt picture character area positioning technology first to determine character area position and character area size, on this technical foundation, adopt character recognition technology identification word content; For the characteristic of picture character features classification, can detect advanced pedestrian's face, carry out again Face Detection after detecting people's face, to determine characteristic.Introduce successively the concrete grammar that the present embodiment preferably extracts characteristic below:
For the picture segmentation method that picture segmentation is become to subgraph: the shade of gray difference of first calculating each row (column) neighbor pixel of picture, according to pixel grey scale change trend judgement sharp point, then line segment frontier point being formed is analyzed, judge whether these boundary sections form the specific region satisfying condition, if, determine that this region is image boundary, and then carry out picture segmentation according to this specific region, while cutting apart, can take the mode that straight line intersects in length and breadth to cut apart, be conducive to like this carry out computerize processing, also can take other modes to cut apart.It should be noted that when a width picture only and while comprising an image, the border searching out is picture border, when a picture comprises multiple image, by said method, can obtain a plurality of image-regions, these image-regions form " subgraph " of pictures.
Extracting method for picture marking area, comprising:
(1) pixel number of every kind of color in the default color set of statistics within the scope of described subgraph, default color set is a preassigned color space, such as being Lab space, set C={c for the color elements that color set comprises
1, c
2..., c
nrepresent, the element number of this set is that the value of N can be hundreds of kind color conventionally;
(2) calculate according to the following formula the significance value of every kind of color:
In formula: c
ifor i element in default color set, the element number that N is color set, w
ifor the pixel number of the i kind color in default color set within the scope of subgraph, dist (c
kc
i) be the distance between two kinds of colors in color set;
(3) in subgraph, the color significance value of each pixel is as the significance value of this pixel, and the region that pixel significance value sum is greater than to predetermined threshold value is identified as foreground area;
(4) can know thus the information such as foreground color number, foreground area position, foreground area number after determining foreground area according to above-mentioned steps.
For the localization method in picture Chinese word region, the present embodiment provides following two kinds of concrete method for optimizing:
The one, based on texture statistics method, the method catches word to have unique texture features, for example word has strong edge, word region gray-value variation is larger, background color before word region has obviously, word region is above has blank with following part, utilize these features to judge by sorter whether certain region is character area, and its step comprises: (1) with the moving window scan image of different scale to obtain a plurality of candidate window; (2) for each candidate window, calculate the textural characteristics of pixel in this candidate window, the modes such as textural characteristics can be by calculating gray scale difference value, extract edge feature, carry out discrete cosine transform (DCT), wavelet transformation, Gabor conversion obtain; (3) by sorter, judge whether each candidate window is character area, and then realize the location of character area.
The 2nd, based on region analysis method, the method catches word to have obvious provincial characteristics, such as most of word is all sealing, region independently, or because of the close formation connected component of pixel color, or drawn a circle to approve in a closed region by a strong edge, its step comprises: (1) extracts all connected components in image, and leaching process can adopt the methods such as binarization of gray value, color cluster, rim detection to realize; (2) utilize the characteristic information of connected component to filter non-legible region, these characteristic informations comprise geological information (as area, length breadth ratio, numbers of hole), fitted ellipse information, stroke information, bone information, colour consistency etc.; (3) adopt bottom-up method that discrete word connected component is merged, determine the border of character area, thereby realize character area location and obtain character area size.
By abovementioned steps, obtain after the character area in image, can further for this character area, realize word and extract, word extracts and can adopt KD tree method of identification to realize, and the method comprises: set KD tree identification rudimentary model; Word in traversal character area, trains the word of first identification, to revise KD tree model of cognition, by this model, again identifies until identify the whole words in character area.In order to improve recognition effect, then can also carry out pre-service to character area before identification, pre-service comprises that typesetting analysis, binaryzation, text line detect, cut word etc.
For method for detecting human face, the present embodiment preferably carries out in accordance with the following steps: first by a large amount of people's face picture of artificial mark and non-face picture, for these people's faces and non-face picture, carry out quasi-Haar wavelet feature extraction, then according to the quasi-Haar wavelet characteristic use Adboost extracting, learn out differentiation people face and non-face level sorter; With different scale moving window scan image, obtain candidate's subwindow; Whether be people face, finally the candidate window of people's face merges when judging, and obtains final human face region if utilizing level sorter to differentiate each candidate window.After determining human face region, can determine accordingly that people's face number, people's face position and people in image are bold little.For Face Detection, the present embodiment preferably carries out in accordance with the following steps: first by a large amount of colour of skin picture of artificial mark, learn out the gauss hybrid models of skin color for these colour of skin pictures; According to each pixel in colour of skin gauss hybrid models computed image, be the probability of the colour of skin to obtain skin color probability distribution plan, the probability of pixel and setting threshold are compared, over the region of this threshold value, be the area of skin color with the specific colour of skin.
Step S204: whether the characteristic that judgement is extracted mates with described feature classification characteristic of correspondence data, if coupling performs step S205; If do not mated, perform step S206;
Here the characteristic of extracting may concrete mode there are differences for different characteristics from the coupling between default characteristic, such as, characteristic for background complexity, it is numeric type comparison, if complexity is 100 minutes systems, as long as the complexity value of the characteristic of extracting operates pre-set threshold value, (default characteristic) thought and can be realized coupling; Characteristic for background colour, it is the comparison of two-valued logic type, if default characteristic background colour is white pure color,, as long as the background colour of the characteristic of extracting is that its allochromatic colour (comprising non-white pure color, tertiary colour etc.) is thought and can not be realized coupling, the matching process of other characteristics is similar.
Step S205: be a class by picture merger to be sorted; These pictures that are classified as a class can be stored, so that down-stream operates on it.
Step S206: judge whether picture to be sorted is disposed, if not, return to step S202; If so, finish picture classification flow process.
The present embodiment is treated category images according to picture background, picture prospect, picture character and four feature classifications of picture personage and has been carried out mechanized classification.Fig. 3 shows the design sketch of the present embodiment, wherein: Fig. 3 (a) is the former figure of picture to be sorted; Fig. 3 (b) is feature project and characteristic, having listed picture background feature class object characteristic background colour is herein that white solid background, picture prospect position are picture centre position, picture character feature class object characteristic is commodity LOGO (O.S.A), in the upper left corner, the characteristic of picture character features classification is for there being model; Fig. 3 (c) is the marking area figure extracting, and translucent grid area is background; Fig. 3 (d) is the character area figure extracting, and translucent yellow area is character area; Fig. 3 (e) is people's face location map, the region behaviour face in blue frame.
It should be noted that, when reading feature classification, read the present embodiment a plurality of feature classifications simultaneously, and a plurality of characteristics in the setting of feature class object characteristic, have also been set, in fact, in industry application, these feature classifications and characteristic independently can be carried out respectively, even, on the present embodiment basis, increase more feature classification and characteristic.The requirement to the accuracy of picture classification and efficiency is all depended in selection under these various situations, conventionally, feature classification and characteristic are more, the accuracy of picture classification is higher, but because the workload of extraction and matching process increases, the efficiency of picture classification will be affected again, therefore, comprehensive, need the needs of this two aspect of balance to come actual selected characteristic classification and characteristic.
Further, the present embodiment has carried out mechanized classification to picture.On the basis of picture classification, can carry out the processing of picture, such as, the processing such as the sign that unitizes on the picture of classification stack, picture cutting, the adjustment of picture pixel.Specifically carry out which kind of type of process and how to process and can set by default processing rule, such as, need to unify the size dimension of the picture of a classification now, can set and process type for " cutting ", processing rule can be set the width, length of cutting etc.In concrete application, can also will preset processing rule and feature classification connects, be certain default processing rule corresponding to feature classification, such as, feature classification is picture background if, and default processing rule can be that the picture background of the picture of a classification is strengthened or weaken processing, to adapt to the needs of certain light intensity, also such as, feature classification is picture character, default processing rule can be the word in picture to be carried out to the adjustment of an order, size, color etc.
To stamp the unified example that is designated in the precalculated position of picture: because picture is classified according to pre-provisioning request, the operation of unified marking is become simply, fast, only need mechanical type to repeat marking, and without worrying whether this operation hinders other picture element.Therefore,, with respect to the mass picture of non-classified, the efficiency of this processing procedure can improve greatly.Picture through unified stamped mark can conveniently be shown, manage picture, and especially, in commercial operation platform, the unitized displaying to commodity (such as clothes, knapsack), can strengthen brand names image, improves clicking rate and the buying rate of commodity.Referring to Fig. 3 (f), the figure shows and through the lower right corner of sorted picture, stamping the effect after sign (Tmall.com).
Embodiment tri-
Above embodiment describes the application's method in detail, and correspondingly, the application also provides a kind of picture apparatus for automatically sorting.Referring to Fig. 4, the figure shows the structured flowchart of picture mechanized classification device of the application's embodiment tri-.This device embodiment 400 comprises: receiving element 401, reading unit 402, extraction unit 403, matching unit 404 and classification unit 405, wherein:
Receiving element 401, for receiving picture to be sorted;
Reading unit 402, for reading the feature classification in feature database;
Sorting out unit 405, is a class for the picture merger to be sorted that characteristic can be mated.
The course of work of this device embodiment 400 is: receiving element 401 receives picture to be sorted and reading unit 402 reads after the feature classification in feature database, by extraction unit 403, according to the described feature classification reading, extracted the characteristic of described picture to be sorted, then, matching unit 404 carries out matching operation by the picture feature data of the extraction default characteristic corresponding with described feature classification, after overmatching, sorting out the picture merger to be sorted that unit 405 can be mated by characteristic is a class.
This device embodiment is receiving picture to be sorted and is extracting after the characteristic of picture to be sorted according to the feature classification in feature database, the default characteristic that these characteristics are corresponding with feature classification is mated, if can mate, by its merger, be a class.Compared with prior art, this device embodiment can automatic acquisition feature classification and with feature classification characteristic of correspondence data, and the characteristic of acquisition is mated to distinguish classification with the characteristic of extraction, thereby can automatically complete the classification of mass picture, save human and material resources and financial cost, improved the work efficiency of picture classification.And, owing to taking robotization treatment measures, as long as pre-set feature classification and characteristic, can by one by one coupling realize picture classification, thereby improved the accuracy of picture classification.The picture classification mode of this device embodiment high-level efficiency and pin-point accuracy has met the application demand of various occasions preferably.
It should be noted that the receiving element 401 of said apparatus embodiment and reading unit 402 are according to the needs of practical application, the operation of reading unit 402 is carried out in the operation that can first carry out receiving element 401 again, also can be contrary, and even both carry out simultaneously.The situation that the first executable operations of receiving element is mainly applicable to carry out for different pictures to be sorted personalized classification, like this, user can set different feature classification and characteristic to the picture to be sorted of some according to the first impression for the treatment of category images, and this is applicable to a small amount of picture to carry out " essence classification "; The first executable operations of reading unit is mainly applicable to unify for the picture to be sorted of magnanimity the situation of classification, so only need to carry out read operation one time, each follow-up picture to be sorted all adopts this feature classification and characteristic, and this is suitable for mass picture to carry out " just classification ".The application person of this device can have an execution sequence that application scenarios is selected above-mentioned associative cell according to various.
The feature classification data that extraction unit 403 in said apparatus embodiment is being processed are different, can have different internal structure.For fear of repetition, under regard to the technique effect that extraction unit different internal structure brings can be referring to the description of aforementioned manner embodiment appropriate section.
(1) when feature classification is picture background and/or picture prospect, described and feature classification characteristic of correspondence data comprise background color number, the background colour of picture background, and/or, when the foreground color number of picture prospect, foreground area number, foreground area position, extraction unit 403 can comprise: divide subelement and extract subelement, wherein: division unit, for described picture to be sorted is divided into one or more subgraphs; Extract subelement, for extract background color number, the background colour of each subgraph according to the described feature classification reading, and/or, foreground color number, foreground area number, the foreground area position of extracting each subgraph.
Further preferably, when extracting subelement when extracting foreground color number, foreground area number, the foreground area position of subgraph, this subelement can also comprise: pixel statistics subelement, significance value computation subunit, foreground area recognin unit and definite subelement, wherein: pixel statistics subelement, for adding up every kind of color in default color set pixel number within the scope of described subgraph; Described significance value computation subunit, for calculating according to the following formula the significance value of every kind of color, and in subgraph the color significance value of each pixel as the significance value of this pixel:
In formula: c
ifor i element in default color set, the element number that N is color set, w
ifor the pixel number of the i kind color in default color set within the scope of subgraph, dist (c
kc
i) be the distance between two kinds of colors in color set;
Foreground area recognin unit, is identified as foreground area for pixel significance value sum being greater than to the region of predetermined threshold value; Determine subelement, for determining foreground color number, foreground area number, foreground area position according to the foreground area of identification.
(2) when feature classification is picture character, when described and feature classification characteristic of correspondence data comprise character area position, character area size and/or word content, extraction unit 403 can comprise: locator unit, and/or, recognin unit, wherein: locator unit, for described picture to be sorted is carried out to character area location, to determine character area position and/or character area size; Recognin unit, locates laggard style of writing word identification for described picture to be sorted being carried out to character area, to determine the word content of character area.Further preferably, locator unit is by carrying out character area location based on texture statistics method or based on region analysis method, and/or described recognin unit is set method of identification by KD the word in character area is identified.
Embodiment tetra-
Embodiment tri-has realized the mechanized classification to picture.In most of the cases, picture is classified and is not final purpose, on basis of classification, also need to carry out picture processing.Picture processing comprises various types of other processing, such as, the sign that unitizes on the picture of classification stack, picture cutting, the adjustment of picture pixel etc.The present embodiment was carried a kind of device that carries out pictorial information interpolation on category images basis.Referring to Fig. 5, the figure shows the composition frame chart of the picture processing device based on picture classification of the embodiment of the present application four.This device embodiment 500 comprises: receiving element 501, reading unit 502, extraction unit 503, matching unit 504, classification unit 505 and processing unit 506, wherein:
Receiving element 501, for receiving picture to be sorted;
Reading unit 502, for reading the feature classification in feature database;
Sorting out unit 505, is a class for the picture merger to be sorted that characteristic can be mated;
The course of work of this device embodiment 500 is: receiving element 501 receives picture to be sorted and reading unit 502 reads after the feature classification in feature database, by extraction unit 503, according to the described feature classification reading, extracted the characteristic of described picture to be sorted, then, matching unit 504 carries out matching operation by the picture feature data of the extraction default characteristic corresponding with described feature classification, after overmatching, sorting out the picture merger to be sorted that unit 505 can be mated by characteristic is a class, then by processing unit, at the predeterminated position of the picture of a classification, add presupposed information.
This device embodiment is before processing picture, owing to having taked picture classification measure, make the picture of processing there are relatively uniform form, size etc., form inherent consistance feature, thereby the batch that is conducive to carry out picture is processed rapidly, has improved the treatment effeciency of picture.
It should be noted that, in the present embodiment except picture processing unit, other unit can be integrated into picture classification device, and then the various concrete application of foundation based on this picture classification device, and picture classification device also can have the various concrete structures of embodiment tri-, to meet the processing requirements to different pieces of information object.Such as, commercialized running platform need to be shown the picture of a large amount of clothes of taking on website.Due to the factor of the clothes of taking itself (such as, clothes classification, color, size etc.) and the impact of the each side condition such as photographed scene (such as daytime, evening etc.), clap look picture out and differ larger, at this moment can adopt the picture classification device that the application provides these pictures to be preset to the classification of feature classification and characteristic, and then will be through the sorted picture presentation of picture classification device out by display unit.Before displaying, also can on the predeterminated position of other picture of same class, add by processing unit commodity LOGO information.So not only improve picture presentation efficiency, and strengthened the brand image of enterprise.
Each embodiment in this instructions all adopts the mode of going forward one by one to describe, between each embodiment identical similar part mutually referring to, each embodiment stresses is the difference with other embodiment.Especially, for system embodiment, because it is substantially similar in appearance to embodiment of the method, so describe fairly simplely, relevant part is referring to the part explanation of embodiment of the method.System embodiment described above is only schematic, the wherein said unit as separating component explanation can or can not be also physically to separate, the parts that show as unit can be or can not be also physical locations, can be positioned at a place, or also can be distributed in a plurality of network element.Can select according to the actual needs some or all of module wherein to realize the object of the present embodiment scheme.Those of ordinary skills, in the situation that not paying creative work, are appreciated that and implement.
The application can be used in numerous general or special purpose computingasystem environment or configuration.For example: personal computer, server computer, handheld device or portable set, plate equipment, multicomputer system, the system based on microprocessor, set top box, programmable consumer-elcetronics devices, network PC, small-size computer, mainframe computer, comprise distributed computing environment of above any system or equipment etc.
The application can describe in the general context of the computer executable instructions of being carried out by computing machine, for example program module.Usually, program module comprises the routine carrying out particular task or realize particular abstract data type, program, object, assembly, data structure etc.Also can in distributed computing environment, put into practice the application, in these distributed computing environment, by the teleprocessing equipment being connected by communication network, be executed the task.In distributed computing environment, program module can be arranged in the local and remote computer-readable storage medium that comprises memory device.
The above is only the application's embodiment; it should be pointed out that for those skilled in the art, do not departing under the prerequisite of the application's principle; can also make some improvements and modifications, these improvements and modifications also should be considered as the application's protection domain.
Claims (16)
1. picture automatic partition class methods, is characterized in that, the method comprises:
Receive picture to be sorted;
Read the feature classification in feature database;
According to described feature classification, extract the characteristic of described picture to be sorted;
The characteristic default characteristic corresponding with described feature classification that coupling is extracted, the picture merger to be sorted that characteristic can be mated is a class.
2. method according to claim 1, is characterized in that, described feature classification comprises picture background, picture prospect, picture character and/or picture personage.
3. method according to claim 2, it is characterized in that, described feature classification is picture background and/or picture prospect, the described default characteristic corresponding with feature classification comprises background color number, the background colour of picture background, and/or, the foreground color number of picture prospect, foreground area number, foreground area position: the described characteristic according to the described picture to be sorted of feature classification extraction comprises:
Described picture to be sorted is divided into one or more subgraphs;
Extract background color number, the background colour of each subgraph, and/or, foreground color number, foreground area number, the foreground area position of extracting each subgraph.
4. method according to claim 3, is characterized in that, foreground color number, foreground area number, the foreground area position of extracting subgraph specifically comprise:
The pixel number of every kind of color in the default color set of statistics within the scope of described subgraph;
Calculate according to the following formula the significance value of every kind of color:
In formula: c
ifor i element in default color set, the element number that N is color set, w
ifor the pixel number of the i kind color in default color set within the scope of subgraph, dist (c
kc
i) be the distance between two kinds of colors in color set;
In subgraph, the color significance value of each pixel is as the significance value of this pixel;
The region that pixel significance value sum is greater than to predetermined threshold value is identified as foreground area;
According to the foreground area of identification, determine foreground color number, foreground area number, foreground area position.
5. method according to claim 2, it is characterized in that, described feature classification is picture character, the described default characteristic corresponding with feature classification comprises character area position, character area size and/or word content: the described characteristic according to the described picture to be sorted of feature classification extraction comprises:
Described picture to be sorted is carried out to character area location, to determine character area position and/or character area size; And/or,
Described picture to be sorted is carried out to character area and locate laggard style of writing word identification, to determine the word content of character area.
6. method according to claim 5, is characterized in that, described picture to be sorted is carried out to character area location and specifically comprise: by carrying out character area location based on texture statistics method or based on region analysis method, and/or,
Described picture to be sorted is carried out to word identification specifically to be comprised: by KD, set method of identification the word in character area is identified.
7. method according to claim 2, it is characterized in that, described feature classification is picture personage, the described default characteristic corresponding with feature classification comprises people's face number, people's face position and/or face complexion: the described characteristic of extracting described picture to be sorted according to feature classification is for extracting people's face number, people's face position and/or the face complexion of described picture to be sorted.
8. the image processing method based on picture classification, is characterized in that, the method comprises:
Receive picture to be sorted;
Read the feature classification in feature database;
According to described feature classification, extract the characteristic of described picture to be sorted;
The characteristic default characteristic corresponding with described feature classification that coupling is extracted, the picture merger to be sorted that characteristic can be mated is a class;
The picture of a classification is processed according to default processing rule.
9. method according to claim 8, is characterized in that, the feature classification in described default processing rule and feature database has corresponding relation, described the picture of a classification is processed and is specially according to default processing rule:
The picture of a classification is processed according to default processing rule corresponding to the feature classification with reading from feature database.
10. method according to claim 8, is characterized in that, described the picture of a classification is processed and is specially according to default processing rule:
Predeterminated position at the picture of a classification adds presupposed information.
11. 1 kinds of picture mechanized classification devices, is characterized in that, this device comprises: receiving element, reading unit, extraction unit, matching unit and classification unit, wherein:
Described receiving element, for receiving picture to be sorted;
Described reading unit, for reading the feature classification in feature database;
Described extraction unit, for extracting the characteristic of described picture to be sorted according to the described feature classification reading;
Described matching unit, for mating the picture feature data default characteristic corresponding with described feature classification of extraction;
Described classification unit is a class for the picture merger to be sorted that characteristic can be mated.
12. devices according to claim 11, it is characterized in that, described feature classification is picture background and/or picture prospect, the described default characteristic corresponding with feature classification comprises background color number, the background colour of picture background, and/or, the foreground color number of picture prospect, foreground area number, foreground area position, described extraction unit comprises: divide subelement and extract subelement, wherein:
Described division unit, for being divided into one or more subgraphs by described picture to be sorted;
Described extraction subelement, for extract background color number, the background colour of each subgraph according to the described feature classification reading, and/or, foreground color number, foreground area number, the foreground area position of extracting each subgraph.
13. devices according to claim 12, it is characterized in that, described extraction subelement is when extracting foreground color number, foreground area number, the foreground area position of subgraph, this subelement comprises: pixel statistics subelement, significance value computation subunit, foreground area recognin unit and definite subelement, wherein:
Described pixel statistics subelement, for adding up every kind of color in default color set pixel number within the scope of described subgraph;
Described significance value computation subunit, for calculating according to the following formula the significance value of every kind of color, and in subgraph the color significance value of each pixel as the significance value of this pixel:
In formula: c
ifor i element in default color set, the element number that N is color set, w
ifor the pixel number of the i kind color in default color set within the scope of subgraph, dist (c
kc
i) be the distance between two kinds of colors in color set;
Described foreground area recognin unit, is identified as foreground area for pixel significance value sum being greater than to the region of predetermined threshold value;
Described definite subelement, for determining foreground color number, foreground area number, foreground area position according to the foreground area of identification.
14. devices according to claim 11, it is characterized in that, described feature classification is picture character, the described default characteristic corresponding with feature classification comprises character area position, character area size and/or word content, described extraction unit comprises: locator unit, and/or, recognin unit, wherein:
Described locator unit, for described picture to be sorted is carried out to character area location, to determine character area position and/or character area size;
Described recognin unit, locates laggard style of writing word identification for described picture to be sorted being carried out to character area, to determine the word content of character area.
15. devices according to claim 14, it is characterized in that, described locator unit is by carrying out character area location based on texture statistics method or based on region analysis method, and/or described recognin unit is set method of identification by KD the word in character area is identified.
16. 1 kinds of picture processing devices based on picture classification, is characterized in that, this device comprises: receiving element, reading unit, extraction unit, matching unit, classification unit and processing unit, wherein:
Described receiving element, for receiving picture to be sorted;
Described reading unit, for reading the feature classification in feature database;
Described extraction unit, for extracting the characteristic of described picture to be sorted according to the described feature classification reading;
Described matching unit, for mating the picture feature data default characteristic corresponding with described feature classification of extraction;
Described classification unit is a class for the picture merger to be sorted that characteristic can be mated;
Described processing unit, for processing according to default processing rule the picture of a classification.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210276320.0A CN103577475B (en) | 2012-08-03 | 2012-08-03 | A kind of picture mechanized classification method, image processing method and its device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210276320.0A CN103577475B (en) | 2012-08-03 | 2012-08-03 | A kind of picture mechanized classification method, image processing method and its device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103577475A true CN103577475A (en) | 2014-02-12 |
| CN103577475B CN103577475B (en) | 2018-01-30 |
Family
ID=50049273
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210276320.0A Active CN103577475B (en) | 2012-08-03 | 2012-08-03 | A kind of picture mechanized classification method, image processing method and its device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103577475B (en) |
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104331490A (en) * | 2014-11-14 | 2015-02-04 | 北京国双科技有限公司 | Network data processing method and device |
| CN104915344A (en) * | 2014-03-10 | 2015-09-16 | 联想(北京)有限公司 | Information processing method and electronic equipment |
| WO2015139204A1 (en) * | 2014-03-18 | 2015-09-24 | 华为终端有限公司 | Picture management method and device |
| CN105120334A (en) * | 2015-08-17 | 2015-12-02 | 无锡天脉聚源传媒科技有限公司 | A method and apparatus for processing television program pictures |
| CN105120335A (en) * | 2015-08-17 | 2015-12-02 | 无锡天脉聚源传媒科技有限公司 | A method and apparatus for processing television program pictures |
| CN105245440A (en) * | 2015-10-23 | 2016-01-13 | 努比亚技术有限公司 | Method, system and operation method for pushing picture messages |
| WO2016066042A1 (en) * | 2014-10-29 | 2016-05-06 | 阿里巴巴集团控股有限公司 | Segmentation method for commodity picture and device thereof |
| CN105843827A (en) * | 2015-05-27 | 2016-08-10 | 维沃移动通信有限公司 | Picture classification method and device for mobile terminal |
| CN105868680A (en) * | 2015-11-24 | 2016-08-17 | 乐视致新电子科技（天津）有限公司 | Channel logo classification method and apparatus |
| CN106021573A (en) * | 2016-05-31 | 2016-10-12 | 北京小米移动软件有限公司 | Image classification method and device and electronic device |
| CN106547893A (en) * | 2016-11-03 | 2017-03-29 | 福建中金在线信息科技有限公司 | A kind of photo sort management system and photo sort management method |
| CN106599940A (en) * | 2016-11-25 | 2017-04-26 | 东软集团股份有限公司 | Picture character identification method and apparatus thereof |
| CN106682683A (en) * | 2016-11-03 | 2017-05-17 | 知酒（上海）网络科技有限公司 | Wine label picture identification method and device |
| CN106886789A (en) * | 2015-12-16 | 2017-06-23 | 芋头科技（杭州）有限公司 | A kind of image recognition sorter and method |
| CN107016392A (en) * | 2016-01-27 | 2017-08-04 | 四川效率源信息安全技术股份有限公司 | A kind of method of text border in removal picture |
| CN107066571A (en) * | 2017-04-07 | 2017-08-18 | 深圳市金立通信设备有限公司 | A kind of image processing method and its equipment |
| CN107944022A (en) * | 2017-12-11 | 2018-04-20 | 努比亚技术有限公司 | Picture classification method, mobile terminal and computer-readable recording medium |
| CN108595661A (en) * | 2018-04-28 | 2018-09-28 | 珠海格力电器股份有限公司 | Method, system, storage medium, photographing device and the mobile phone of photo disposal |
| CN108717530A (en) * | 2018-05-21 | 2018-10-30 | Oppo广东移动通信有限公司 | Image processing method, device, computer readable storage medium and electronic equipment |
| CN110222207A (en) * | 2019-05-24 | 2019-09-10 | 珠海格力电器股份有限公司 | Method for sorting, device and the intelligent terminal of picture |
| CN110362702A (en) * | 2019-05-27 | 2019-10-22 | 中国联合网络通信集团有限公司 | Picture management method and equipment |
| CN110097082B (en) * | 2019-03-29 | 2021-08-27 | 广州思德医疗科技有限公司 | Splitting method and device of training set |
| CN116912721A (en) * | 2023-09-14 | 2023-10-20 | 众芯汉创(江苏)科技有限公司 | Power distribution network equipment body identification method and system based on monocular stereoscopic vision |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020097892A1 (en) * | 1998-06-01 | 2002-07-25 | Nec Corporation | Digital watermark insertion system and digital watermark characteristic table creating device |
| CA2433307C (en) * | 2002-07-01 | 2007-01-02 | Xerox Corporation | Segmentation method and system for multiple raster content (mrc) representation of documents |
| CN101055592A (en) * | 2007-05-20 | 2007-10-17 | 宁尚国 | Image information generation method and device |
| CN101635763A (en) * | 2008-07-23 | 2010-01-27 | 深圳富泰宏精密工业有限公司 | Picture classification system and method |
| CN102165487A (en) * | 2008-09-25 | 2011-08-24 | 乐天株式会社 | Foreground area extracting program, foreground area extracting apparatus and foreground area extracting method |
-
2012
- 2012-08-03 CN CN201210276320.0A patent/CN103577475B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020097892A1 (en) * | 1998-06-01 | 2002-07-25 | Nec Corporation | Digital watermark insertion system and digital watermark characteristic table creating device |
| CA2433307C (en) * | 2002-07-01 | 2007-01-02 | Xerox Corporation | Segmentation method and system for multiple raster content (mrc) representation of documents |
| CN101055592A (en) * | 2007-05-20 | 2007-10-17 | 宁尚国 | Image information generation method and device |
| CN101635763A (en) * | 2008-07-23 | 2010-01-27 | 深圳富泰宏精密工业有限公司 | Picture classification system and method |
| CN102165487A (en) * | 2008-09-25 | 2011-08-24 | 乐天株式会社 | Foreground area extracting program, foreground area extracting apparatus and foreground area extracting method |
Non-Patent Citations (6)
| Title |
|---|
| 崔宁海等: "空间主颜色描述符的图像特征提取算法", 《沈阳理工大学学报》 * |
| 杨楠: "基于内容的商品图像分类技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 杨育彬等: "一种基于颜色连通的图像纹理检索新方法", 《电子学报》 * |
| 林克正等: "基于分块主颜色匹配的图像检索", 《计算机工程》 * |
| 潘宇: "票据图像分类的技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 赵倩等: "结合高斯多尺度变换和颜色复杂度计算的显著区域检测", 《仪器仪表学报》 * |
Cited By (31)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104915344A (en) * | 2014-03-10 | 2015-09-16 | 联想(北京)有限公司 | Information processing method and electronic equipment |
| CN104915344B (en) * | 2014-03-10 | 2019-04-26 | 联想(北京)有限公司 | A kind of information processing method and electronic equipment |
| WO2015139204A1 (en) * | 2014-03-18 | 2015-09-24 | 华为终端有限公司 | Picture management method and device |
| US10297029B2 (en) | 2014-10-29 | 2019-05-21 | Alibaba Group Holding Limited | Method and device for image segmentation |
| WO2016066042A1 (en) * | 2014-10-29 | 2016-05-06 | 阿里巴巴集团控股有限公司 | Segmentation method for commodity picture and device thereof |
| CN104331490B (en) * | 2014-11-14 | 2018-07-13 | 北京国双科技有限公司 | network data processing method and device |
| CN104331490A (en) * | 2014-11-14 | 2015-02-04 | 北京国双科技有限公司 | Network data processing method and device |
| CN105843827A (en) * | 2015-05-27 | 2016-08-10 | 维沃移动通信有限公司 | Picture classification method and device for mobile terminal |
| CN105120334B (en) * | 2015-08-17 | 2018-12-21 | 无锡天脉聚源传媒科技有限公司 | A kind of method and apparatus handling TV programme picture |
| CN105120335A (en) * | 2015-08-17 | 2015-12-02 | 无锡天脉聚源传媒科技有限公司 | A method and apparatus for processing television program pictures |
| CN105120334A (en) * | 2015-08-17 | 2015-12-02 | 无锡天脉聚源传媒科技有限公司 | A method and apparatus for processing television program pictures |
| CN105120335B (en) * | 2015-08-17 | 2018-08-24 | 无锡天脉聚源传媒科技有限公司 | A kind of method and apparatus of processing TV programme picture |
| CN105245440A (en) * | 2015-10-23 | 2016-01-13 | 努比亚技术有限公司 | Method, system and operation method for pushing picture messages |
| CN105868680A (en) * | 2015-11-24 | 2016-08-17 | 乐视致新电子科技（天津）有限公司 | Channel logo classification method and apparatus |
| CN106886789A (en) * | 2015-12-16 | 2017-06-23 | 芋头科技（杭州）有限公司 | A kind of image recognition sorter and method |
| CN107016392A (en) * | 2016-01-27 | 2017-08-04 | 四川效率源信息安全技术股份有限公司 | A kind of method of text border in removal picture |
| CN106021573A (en) * | 2016-05-31 | 2016-10-12 | 北京小米移动软件有限公司 | Image classification method and device and electronic device |
| CN106682683A (en) * | 2016-11-03 | 2017-05-17 | 知酒（上海）网络科技有限公司 | Wine label picture identification method and device |
| CN106547893A (en) * | 2016-11-03 | 2017-03-29 | 福建中金在线信息科技有限公司 | A kind of photo sort management system and photo sort management method |
| CN106599940B (en) * | 2016-11-25 | 2020-04-17 | 东软集团股份有限公司 | Picture character recognition method and device |
| CN106599940A (en) * | 2016-11-25 | 2017-04-26 | 东软集团股份有限公司 | Picture character identification method and apparatus thereof |
| CN107066571A (en) * | 2017-04-07 | 2017-08-18 | 深圳市金立通信设备有限公司 | A kind of image processing method and its equipment |
| CN107944022A (en) * | 2017-12-11 | 2018-04-20 | 努比亚技术有限公司 | Picture classification method, mobile terminal and computer-readable recording medium |
| CN108595661A (en) * | 2018-04-28 | 2018-09-28 | 珠海格力电器股份有限公司 | Method, system, storage medium, photographing device and the mobile phone of photo disposal |
| CN108717530A (en) * | 2018-05-21 | 2018-10-30 | Oppo广东移动通信有限公司 | Image processing method, device, computer readable storage medium and electronic equipment |
| CN108717530B (en) * | 2018-05-21 | 2021-06-25 | Oppo广东移动通信有限公司 | Image processing method, image processing device, computer-readable storage medium and electronic equipment |
| CN110097082B (en) * | 2019-03-29 | 2021-08-27 | 广州思德医疗科技有限公司 | Splitting method and device of training set |
| CN110222207A (en) * | 2019-05-24 | 2019-09-10 | 珠海格力电器股份有限公司 | Method for sorting, device and the intelligent terminal of picture |
| CN110362702A (en) * | 2019-05-27 | 2019-10-22 | 中国联合网络通信集团有限公司 | Picture management method and equipment |
| CN116912721A (en) * | 2023-09-14 | 2023-10-20 | 众芯汉创(江苏)科技有限公司 | Power distribution network equipment body identification method and system based on monocular stereoscopic vision |
| CN116912721B (en) * | 2023-09-14 | 2023-12-05 | 众芯汉创(江苏)科技有限公司 | Power distribution network equipment body identification method and system based on monocular stereoscopic vision |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103577475B (en) | 2018-01-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103577475A (en) | Picture automatic sorting method, picture processing method and devices thereof | |
| Zhang et al. | A multilevel point-cluster-based discriminative feature for ALS point cloud classification | |
| Jasiewicz et al. | Landscape similarity, retrieval, and machine mapping of physiographic units | |
| Zhang et al. | Integrating bottom-up classification and top-down feedback for improving urban land-cover and functional-zone mapping | |
| CN102012939B (en) | Method for automatically tagging animation scenes for matching through comprehensively utilizing overall color feature and local invariant features | |
| CN104951554B (en) | It is that landscape shines the method for mixing the verse for meeting its artistic conception | |
| Chen et al. | Building detection in an urban area using lidar data and QuickBird imagery | |
| US20180129658A1 (en) | Color sketch image searching | |
| CN108761451B (en) | Land use change intelligent understanding method based on multi-temporal remote sensing image | |
| Ozdarici-Ok | Automatic detection and delineation of citrus trees from VHR satellite imagery | |
| CN108268527B (en) | A method of detection land use pattern variation | |
| CN103198479A (en) | SAR image segmentation method based on semantic information classification | |
| CN112883926B (en) | Identification method and device for form medical images | |
| CN107330027A (en) | A kind of Weakly supervised depth station caption detection method | |
| Li et al. | A method based on an adaptive radius cylinder model for detecting pole-like objects in mobile laser scanning data | |
| Anichini et al. | Developing the ArchAIDE application: a digital workflow for identifying, organising and sharing archaeological pottery using automated image recognition | |
| Wang et al. | Remote-sensing image retrieval by combining image visual and semantic features | |
| CN104462526A (en) | Multi-user online collaboration rapid vectorization method for high-resolution remote sensing images | |
| Lang et al. | Combined object-based classification and manual interpretation–synergies for a quantitative assessment of parcels and biotopes | |
| CN111144270B (en) | Neural network-based handwritten text integrity evaluation method and evaluation device | |
| CN109657728B (en) | Sample production method and model training method | |
| CN113673528B (en) | Text processing method, text processing device, electronic equipment and readable storage medium | |
| CN109492573A (en) | A kind of pointer read method and device | |
| CN106709437A (en) | Improved intelligent processing method for image-text information of scanning copy of early patent documents | |
| CN105760421A (en) | Land-use type classifying method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 1193181 Country of ref document: HK |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: GR Ref document number: 1193181 Country of ref document: HK |