WO2023272659A1

WO2023272659A1 - Method and apparatus for recognizing cover image, storage medium, and recognition device

Info

Publication number: WO2023272659A1
Application number: PCT/CN2021/103865
Authority: WO
Inventors: 张明云
Original assignee: 东莞市小精灵教育软件有限公司
Priority date: 2021-06-30
Filing date: 2021-06-30
Publication date: 2023-01-05

Abstract

Disclosed in embodiments of the present invention are a method and apparatus for recognizing a cover image, a storage medium, and a recognition device. The method comprises: acquiring a cover image to be recognized, a first gallery, and a second gallery; searching the first gallery for a sample cover image that matches the cover image to be recognized; determining whether the sample cover image found in the first gallery is associated with another sample cover; if so, acquiring block images of the cover image to be recognized, searching the second gallery for a predetermined area image the same as a block image, and recognizing the cover image to be recognized as a sample cover image found in the second gallery and corresponding to the predetermined area image; if not, recognizing the cover image to be to be recognized as a sample cover image found in the first gallery. The present application can reduce the probability of mistaken recognition and also reduce the probability of a situation in which page-turning of a picture book cannot be continued and a child cannot continue to read the picture book.

Description

Cover image recognition method, device, storage medium and recognition equipment

technical field

The present application relates to the technical field of data processing, and in particular to a cover image recognition method, device, storage medium and recognition equipment.

Background technique

With the development of science and technology, some electronic devices can now recognize the covers of some reading materials, such as the covers of picture books. After reading the covers of picture books, electronic devices can display the text information of picture books in the form of electronic version, which is realized by audio The page-reading function of picture books reduces the difficulty for children to read picture books and reduces the time for parents to guide and assist children in reading picture books.

technical problem

However, some of the current reading materials are related, such as the first volume and the second volume of the reading material. There are few differences between the cover images of the two reading materials. The difference is often only the words "first volume" and "second volume", so they are related. The readings have a high similarity. When the electronic device recognizes the cover of the reading, it recognizes it according to the similarity between the cover image of the reading in the database and the current cover image of the reading. The page turning of the reading material cannot be continued, and the user cannot continue to read the reading material.

technical solution

The first aspect of the present invention provides a method for identifying a cover image, the method comprising: acquiring a cover image to be identified, a first gallery and a second gallery, the first gallery contains all sample cover images that have been collected, and the second gallery includes The second gallery contains at least one set of predetermined area images of sample cover images, and the sample cover images corresponding to each set of predetermined area images have a correlation, and the sample cover images with the correlation include a series of sample covers The image, or, includes a set of sample cover images in multiple languages; search for a sample cover image that matches the cover image to be identified in the first gallery; determine whether the sample cover image searched in the first gallery is compatible with Other sample cover images are relevant; if so, each block image of the cover image to be identified is obtained, the predetermined area image identical to the block image is searched in the second gallery, and the image to be identified is The cover image is identified as the sample cover image corresponding to the predetermined area image found in the second gallery; if not, the cover image to be identified is identified as the sample cover searched in the first gallery image.

Wherein, the image of the predetermined area contained in the second gallery also has a first label of length and width; the obtaining of each block image of the cover image to be recognized includes: obtaining each block of the cover image to be recognized block position; using the length and width included in the first tag, intercepting each block of the cover image to be recognized to generate a block image.

Wherein, the predetermined area image contained in the second gallery also has a second label, and the second label includes the first image characteristic information and the first text characteristic information of the predetermined area image; the method further includes: When obtaining each block image of the cover image to be recognized, the second image feature information and the second text feature information of the block image are also obtained; the searching in the second gallery is related to the block image the same predetermined area image, and identifying the cover image to be identified as a sample cover image corresponding to the predetermined area image searched in the second gallery includes: searching for the block in the second gallery The first image characteristic information of the image and the predetermined area image represented by the second label with the same first text feature information; identifying the cover image to be recognized as a sample cover image corresponding to the predetermined area image searched in the second gallery .

Wherein, the method also includes: pre-establishing a second gallery; the establishment method of the second gallery includes: acquiring the first gallery, and generating a blank gallery; screening samples with the same correlation in the first gallery Cover image; take a screenshot of the predetermined area on the screened sample cover image to obtain the predetermined area image; obtain the length and width of the predetermined area image, and use the length and width to generate the first label of the predetermined area image; extract the predetermined area image The first image feature information and the first text feature information of the predetermined area image, and use the first image feature information and the first text feature information to generate the second label of the predetermined area image; The predetermined area image of the second label is allocated in the predetermined area of the blank gallery to generate a second gallery. In each predetermined area of the second gallery, the correlation between the sample cover images of each predetermined area is the same, and different predetermined areas Between , the correlation between the sample cover images is different.

Wherein, the method further includes: generating a pre-search result for the sample cover image searched in the first gallery in advance, and performing a search on the pre-search result for the sample cover image searched in the first gallery and other sample cover images Marking; the judging whether the sample cover image searched in the first gallery is related to other sample cover images includes: obtaining the sample cover image that matches the cover image to be identified by searching in the first gallery , to obtain the same search result as the pre-search result; identify whether there is a mark in the pre-search result corresponding to the search result; if so, determine that the sample cover image searched in the first gallery is associated with other sample cover images sex.

Wherein, the cover image to be identified is the cover image of a picture book; the method for obtaining the first gallery and the second gallery includes: reading local storage, or/and, reading cloud storage, obtaining the first gallery and the second gallery Describe the second gallery.

Wherein, both the predetermined area image and the block image include at least the name and serial number of the corresponding sample cover image.

The second aspect of the present invention provides a cover image recognition device, including: a data acquisition module, used to obtain the cover image to be recognized, the first gallery and the second gallery, the first gallery of the cover image to be recognized contains all the collected Sample cover images, the second gallery includes at least a set of predetermined area images of sample cover images, the sample cover images corresponding to each set of predetermined area images have a correlation, and the sample cover images with the correlation , including a series of sample cover images, or, including a set of multilingual sample cover images; a first search module, configured to search the first gallery for sample cover images that match the cover images to be identified; The judging module is used to judge whether the sample cover image searched in the first gallery is related to other sample cover images; the second searching module obtains each block image of the cover image to be identified, Search the same predetermined area image as the block image in the second gallery, and identify the cover image to be identified as a sample cover image corresponding to the predetermined area image searched in the second gallery; an output module for If the judgment of the judging module is negative, the cover image to be recognized is recognized as the sample cover image found in the first gallery.

A third aspect of the present invention provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the processor performs the following steps:

Obtain the cover image to be identified, the first gallery and the second gallery, the first gallery contains all the sample cover images that have been included, the second gallery contains at least one set of predetermined area images of the sample cover images, and each set of predetermined The sample cover images corresponding to the region images have a correlation, and the sample cover images with the correlation include a series of sample cover images, or include a group of multilingual sample cover images; in the first Search for a sample cover image matching the cover image to be identified in the gallery; judge whether the sample cover image searched in the first gallery is associated with other sample cover images; if so, obtain the image of the cover image to be identified For each block image, search for a predetermined area image identical to the block image in the second gallery, and identify the to-be-recognized cover image as the one found in the second gallery and corresponding to the predetermined area image A sample cover image; if not, identifying the to-be-recognized cover image as the sample cover image found in the first gallery.

A fourth aspect of the present invention provides a cover image recognition device, including a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the following steps:

Beneficial effect

Adopting the embodiment of the present invention has the following beneficial effects: firstly, it can be judged whether the searched sample cover image in the first gallery is related to other sample cover images, and the correlation indicates that the searched sample cover image is related to other samples The cover image of the reading material has a higher similarity, so when there is relevance, the present invention searches the second gallery for the cover image to be recognized, and when searching, it uses a block with ten times the cover image The image is matched with the image in the predetermined area in the second gallery, the matching area is reduced, and the difference between the cover image to be recognized and the sample cover image can be searched more accurately, thereby reducing the probability of confusing recognition and reducing the reading time. The probability that the page turning cannot continue and the user cannot continue reading the book.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present application. Those skilled in the art can also obtain other drawings based on these drawings without creative work.

in:

Fig. 1 is the flowchart of the recognition method of cover image in an embodiment;

Fig. 2 is the flow chart of obtaining each block image of the cover image to be identified in the identification method of the cover image in one embodiment;

Fig. 3 is a method for identifying a cover image in an embodiment, searching for an image of a predetermined area identical to a block image in the second gallery, and identifying the cover image to be recognized as the one searched in the second gallery and corresponding to the image of the predetermined area Schematic diagram of the process for a sample cover image;

Fig. 4 is a flowchart of a method for generating a second gallery of a cover image recognition method in an embodiment;

Fig. 5 is a flow chart of judging whether the sample cover image searched in the first gallery is associated with other sample cover images in the cover image recognition method in one embodiment;

Fig. 6 is a schematic block diagram of a cover image recognition device in an embodiment;

Fig. 7 is a schematic block diagram of the structure of a computer device in an embodiment.

Embodiments of the present invention

The following will clearly and completely describe the technical solutions in the embodiments of the application with reference to the drawings in the embodiments of the application. Apparently, the described embodiments are only some of the embodiments of the application, not all of them. Based on the embodiments in this application, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of this application.

Referring to Fig. 1, an embodiment of the present invention provides a method for identifying a cover image, including:

S101. Obtain the cover image to be identified, the first gallery and the second gallery;

S102. Search for a sample cover image matching the cover image to be identified in the first gallery;

S103. Determine whether the sample cover image searched in the first gallery is related to other sample cover images;

S104. If so, obtain each block image of the cover image to be identified, search for an image of a predetermined area identical to the block image in the second gallery, and identify the cover image to be identified as the predetermined area searched in the second gallery The sample cover image corresponding to the image;

S105. If not, identify the cover image to be recognized as a sample cover image found in the first gallery. Wherein, in step S101, the cover image to be identified is the cover image of the picture book, the first gallery contains sample cover images of all picture books that have been included, and the second gallery contains at least one set of predetermined area images of sample cover images, each set of predetermined area The sample cover images corresponding to the images are associated, and the associated sample cover images include a series of sample cover images, or include a group of multilingual sample cover images.

In step S101, the relevant sample cover image may be a sample cover image of a series of picture books, such as the cover images of the first volume, the second volume, and the third volume of the picture book, and for example the first volume and the second volume of the picture book , Volume 2, etc.; a sample cover image of a related picture book, or a set of sample cover images of multilingual picture books, such as the cover image of a Chinese version of a picture book, the cover image of an English version of a picture book, or the Japanese version of a picture book with the same content The cover image, etc.; and the cover image is related, which means that the picture book corresponding to the cover image is related.

In step S102, due to the matching, the picture book corresponding to the sample cover image searched in the first library is most likely to be the same as the picture book corresponding to the cover image to be recognized, so further judgment as in step S103 is required.

In step S103, if it is judged that the sample cover image searched in the first library is not related to other sample cover images, it means that the searched sample cover image does not have a high similarity with other picture book sample cover images Therefore, the identification is correct in this case, and step S105 can be executed directly.

In step S103, if it is determined that the searched sample cover image in the first gallery is related to other sample cover images, it indicates that the searched sample cover image has a high degree of similarity with some other related cover images , so if the search is only performed in the first gallery, there may be confusion and recognition. Therefore, in this case, step S104 is performed to search for the same predetermined area image as the block image in the second gallery, which can more accurately Search for the cover image to be recognized.

In this embodiment, the image of the predetermined area and the block image mentioned above, for example in step S104 , both include at least the name and serial number of the picture book corresponding to the cover image of the picture book.

Therefore, the present invention searches for the cover image to be recognized in the second gallery, and when searching, uses the block image of the real-time cover to match the predetermined area image in the second gallery, reduces the matching area, and can be more accurate The difference between the cover image to be recognized and the sample cover image can be accurately searched, thereby reducing the probability of confusion and recognition, and also reducing the probability that the page turning of the reading material cannot continue, and the user cannot continue reading the reading material.

In one embodiment, the image of the predetermined area contained in the second gallery also has a first label of length and width.

Referring to Fig. 2, in step S104, each block image of the cover image to be recognized is obtained, including:

S1041. Obtain the block position of each block of the cover image to be recognized;

S1042. Using the length and width included in the first tag, intercept each block of the cover image to be recognized to generate a block image.

In this embodiment, the block image contains the picture book name and serial number corresponding to the picture book cover image, and the serial number can be a number or a text, for example, the serial number can be 1, 2, 3, then the picture book name and serial number indicate A specific picture book among multiple related picture books; for example, the serial number can also be Volume 1, Volume 2, Volume 2, etc., and the name and serial number of the picture book can also indicate the specific one among multiple picture books that are related Picture book; for example, the serial number can also be a version number, such as the first edition, second edition, third edition, etc. At this time, the picture book name and serial number can also indicate a specific picture book among the multiple picture books that are related. Therefore, it can be understood that, in the case of the same picture book title, the information that can distinguish different picture books can be used as the serial number in the above.

In step S104, since the position of each block of the cover image of the picture book is generally fixed, at the position above, in the middle, or below the middle of the cover image, according to the length and width contained in the first label, in these to-be Recognize these positions of the cover image and perform multiple screenshots to obtain multiple block images, and each block image is searched in the second gallery to find the block image with the same predetermined area image in the second gallery, then in The picture book cover corresponding to the predetermined area image found in the second gallery is the same as the real-time picture book cover.

In one embodiment, the predetermined area image contained in the second gallery further has a second label, and the second label includes the first image characteristic information and the first text characteristic information of the predetermined area image;

The identification method of the cover image further includes: when obtaining each block image of the cover image to be identified, also obtaining the second image feature information and the second text feature information of the block image;

Please refer to FIG. 3, in step S104, search for the same predetermined area image as the block image in the second gallery, and identify the cover image to be identified as the sample cover image corresponding to the predetermined area image searched in the second gallery ,include:

S1043. Search in the second gallery for a predetermined area image represented by a second label that is identical to the first image characteristic information and the first text characteristic information of the block image;

S1044. Recognize the cover image to be recognized as a sample cover image corresponding to the predetermined area image found in the second gallery.

In step S1043, the second label contains the second image feature information and the second text feature information of the predetermined area image, so if the first image feature information and the first text feature information are the same as the second label, it means that the predetermined area image It is the same as the text information and image information of the block image. In this case, the sample cover image represented by the predetermined area image is the cover image to be recognized represented by the block image.

Please refer to FIG. 4 , in one embodiment, the method for identifying the cover image further includes: establishing a second gallery.

Wherein, the method for establishing the second gallery includes:

S1011. Acquire the first image gallery, and generate a blank image gallery;

S1012. Screen sample cover images with the same relevance in the first gallery;

S1013. Take a screenshot of the predetermined area on the screened sample cover image to obtain an image of the predetermined area;

S1014. Obtain the length and width of the predetermined area image, and use the length and width to generate a first label of the predetermined area image;

S1015. Extract the first image feature information and the first text feature information of the predetermined area image, and use the first image feature information and the first text feature information to generate a second label of the predetermined area image;

S1016. Allocate all the images in the predetermined area with the first label and the second label in the predetermined area of the blank gallery to generate a second gallery.

In step S1016, in each predetermined area of the second gallery, the correlations between the sample cover images in each predetermined area are the same, and the correlations between the sample cover images are different in different predetermined areas.

By making a gallery and distributing the predetermined area images of the associated picture book sample cover images into the predetermined area of the gallery, it is possible to make the picture books corresponding to each predetermined area image in the second gallery have associated picture books, Therefore, when the search is performed in step S104, unrelated picture books will not be searched, thereby reducing the search amount and improving the search speed.

In one embodiment, the method for identifying a cover image further includes: generating pre-search results for the picture books corresponding to the sample cover images searched in the first gallery in advance, and searching for the picture books corresponding to the sample cover images searched in the first gallery and Other picture books are marked with relevant pre-search results.

Please refer to FIG. 5, step S103, to determine whether the picture book corresponding to the sample cover image searched in the first gallery is related to other picture books, including:

S1031. Obtain a sample cover image that matches the cover image to be identified by searching in the first gallery, and obtain the same search result as the pre-search result;

S1032. Identify whether there is a mark in the pre-search result corresponding to the search result;

S1033. If yes, determine that the sample cover image found in the first gallery is related to other sample cover images.

In this embodiment, when the sample cover image in the first gallery is searched out, there will be a search result, so the present invention pre-stores the pre-search results of all sample cover images, and each pre-search result corresponds to a search result. Therefore, the corresponding pre-search result can be obtained through the search result through step S1031.

After the identification in step S1032, if there is a mark in the pre-search result, it means that the picture book corresponding to the searched sample cover image is related to the picture book corresponding to other sample cover images. Correspondingly, if there is no mark in the search result, It means that the picture book corresponding to the searched sample cover image is not associated with other picture books corresponding to the sample cover image, so it can be determined whether the picture book corresponding to the sample cover image searched in the first gallery is related to other picture books.

In one embodiment, the method for acquiring the first gallery and the second gallery includes: reading local storage, or/and, reading cloud storage, and acquiring the first gallery and the second gallery.

Preferably, in this embodiment, both the first gallery and the second gallery are deployed in the cloud, and by requesting data from the cloud, the cloud storage is read to obtain the first gallery and the second gallery. This method can reduce the number of users The memory occupation of the terminal is reduced, and when the first library and the second library need to be updated or maintained, the user does not need to download an update package, thereby facilitating the update and maintenance of the first library and the second library.

In other embodiments, the second gallery and the second gallery can also be deployed in local storage, which can reduce the time for users to request data from the Internet.

The second aspect of the embodiment of the present application provides a cover image recognition device, please refer to Figure 6, the device includes: a data acquisition module 1, a first search module 2, a judgment module 3, a second search module 4 and an output module 5; Obtaining module 1 is used to obtain the cover image to be identified, the first gallery and the second gallery, the first gallery contains all sample cover images that have been included, the second gallery contains at least one set of predetermined area images of sample cover images, and each set of predetermined areas The sample cover images corresponding to the images are related, and the sample cover images with correlation include a series of sample cover images, or include a group of multilingual sample cover images; the first search module 2 is used to search in the first gallery Search for a sample cover image that matches the cover image to be identified; Judgment module 3 is used to judge whether the sample cover image searched out in the first gallery is associated with other sample cover images; If the judgment of 3 is yes, each block image of the cover image to be identified is obtained, the predetermined area image identical to the block image is searched in the second gallery, and the cover image to be identified is identified as searched out in the second gallery, The sample cover image corresponding to the predetermined area image; the output module 5 is used to identify the cover image to be recognized as the sample cover image searched in the first library if the judgment of the judging module 3 is negative.

The cover image identification device provided in this embodiment can first judge whether the picture book corresponding to the sample cover image searched in the first gallery is related to other picture books, and if there is a relationship, it means that the searched sample cover image is related to some other pictures. The cover image of the related picture book has a high similarity, so when there is a relationship, the present invention searches the cover image to be recognized in the second gallery, and when searching, it uses the block image of the real-time cover and The images in the predetermined area in the second gallery are matched, the matching area is reduced, and the difference between the cover image to be recognized and the sample cover image can be searched more accurately, thereby reducing the probability of confusion and recognition, and also reducing the need for reading. Odds that the page will not continue and the user will not be able to continue reading the book.

The second search module 4 includes: a block position acquisition unit and a block image generation unit; the block position acquisition unit is used to obtain the block positions of each block of the cover image to be identified; the block image generation unit is used to use the first The length and width contained in a label, each block of the cover image to be recognized is intercepted, and a block image is generated.

In one embodiment, the image of the predetermined area included in the second gallery further has a second label, and the second label includes the first image characteristic information and the first text characteristic information of the image of the predetermined area.

The identification device of the cover image further includes an information acquisition module for acquiring the second image feature information and the second text feature information of the block images when the second search module 4 acquires each block image of the cover image to be recognized.

In this embodiment, the second search module 4 further includes: a predetermined area image search unit and a cover image recognition unit;

The predetermined area image search unit is used to search the predetermined area image represented by the second label that is the same as the first image characteristic information and the first text characteristic information of the block image in the second gallery;

The cover image recognition unit is configured to identify the cover image to be recognized as a sample cover image corresponding to the predetermined area image searched in the second gallery.

In one embodiment, the data acquisition module 1 includes: a first gallery acquisition unit, a blank gallery generation unit, a screening unit, a screenshot unit, a first label generation unit, a feature information extraction unit, and a second gallery generation unit; the first gallery acquisition The unit is used to obtain the first gallery; the blank gallery generation unit is used to generate a blank gallery; the screening unit is used to screen sample cover images with the same relevance in the first gallery; the screenshot unit is used to pre-order the screened sample cover images A screenshot of the area to obtain an image of the predetermined area; the first label generation unit is used to obtain the length and width of the image of the predetermined area, and uses the length and width to generate the first label of the image of the predetermined area; the feature information extraction unit is used to extract the image of the predetermined area The first image feature information and the first text feature information, and use the first image feature information and the first text feature information to generate the second label of the predetermined area image; the second gallery generation unit is used to use all the first label and the second label. The predetermined area images are distributed in the predetermined areas of the blank gallery, and a second gallery is generated. In each predetermined area of the second gallery, the correlation between the sample cover images in each predetermined area is the same, and between different predetermined areas, the relationship between the sample cover images The correlation between them is different.

In one embodiment, the cover image recognition device further includes: a pre-search result generation and marking module, which is used to generate a pre-search result for the picture book corresponding to the sample cover image searched in the first gallery in advance, and search for the picture book corresponding to the sample cover image in the first gallery. The picture book corresponding to the searched sample cover image is marked with the pre-search results that are related to other picture books.

In this embodiment, the judging module 3 includes: a sample cover image acquisition unit, a mark recognition unit, and a relevance judgment unit; the sample cover image acquisition unit is used to acquire a sample cover that matches the cover image to be identified by searching in the first gallery image, to obtain the same search result as the pre-search result; the tag recognition unit is used to identify whether there is a tag in the pre-search result corresponding to the search result;

The relevance judging unit is configured to determine that the sample cover images searched in the first gallery are related to picture books corresponding to other sample cover images if the tag recognition unit recognizes that the pre-search results corresponding to the search results have tags.

Figure 7 shows a diagram of the internal structure of a computer device in one embodiment. Specifically, the computer device may be a terminal or a server. As shown in FIG. 7, the computer device includes a processor, a memory, and a network interface connected through a system bus. Wherein, the memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium of the computer device stores an operating system, and may also store a computer program. When the computer program is executed by the processor, the processor can realize the method for identifying the cover image. A computer program may also be stored in the internal memory, and when the computer program is executed by the processor, the processor may execute the recognition method of the cover image. Those skilled in the art can understand that the structure shown in Figure 7 is only a block diagram of a part of the structure related to the solution of this application, and does not constitute a limitation to the computer equipment on which the solution of this application is applied. The specific computer equipment can be More or fewer components than shown in the figures may be included, or some components may be combined, or have a different arrangement of components.

The embodiment of the present application also provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by the processor, the processor performs the following steps:

Obtain the cover image to be identified, the first gallery and the second gallery, the first gallery contains all sample cover images that have been included, the second gallery contains at least one set of predetermined area images of sample cover images, and the sample cover corresponding to each set of predetermined area images The images are related, and the sample cover images with relevance include a series of sample cover images, or include a group of multilingual sample cover images; search for sample cover images that match the cover images to be identified in the first gallery ; Judging whether the sample cover image searched out in the first gallery is related to other sample cover images; area image, and identify the cover image to be identified as the sample cover image corresponding to the predetermined area image searched out in the second gallery; if not, identify the cover image to be identified as the sample cover image searched out in the first gallery .

The embodiment of the present application also provides a picture book recognition device, including a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the processor is made to perform the following steps:

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be realized through computer programs to instruct related hardware, and the programs can be stored in a non-volatile computer-readable storage medium When the program is executed, it may include the processes of the embodiments of the above-mentioned methods. Wherein, any references to memory, storage, database or other media used in the various embodiments provided in the present application may include non-volatile and/or volatile memory. Nonvolatile memory can include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Chain Synchlink DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

The technical features of the above embodiments can be combined arbitrarily. To make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be It is considered to be within the range described in this specification.

The above-mentioned embodiments only express several implementation modes of the present application, and the description thereof is relatively specific and detailed, but should not be construed as limiting the patent scope of the present application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the scope of protection of the patent application should be based on the appended claims.

Claims

A method for identifying a cover image, characterized in that the method comprises:

Obtain the cover image to be identified, the first gallery and the second gallery, the first gallery contains all the sample cover images that have been included, the second gallery contains at least one set of predetermined area images of the sample cover images, and each set of predetermined The sample cover images corresponding to the region images have a correlation, and the sample cover images with the correlation include a series of sample cover images, or include a group of multilingual sample cover images;

Searching for a sample cover image matching the cover image to be identified in the first gallery;

Judging whether the sample cover image searched in the first gallery is related to other sample cover images;

If so, obtain each block image of the cover image to be identified, search for an image of a predetermined area identical to the block image in the second gallery, and identify the cover image to be identified as being in the first image The sample cover image corresponding to the image in the predetermined area searched out of the second library;

If not, the to-be-recognized cover image is identified as the sample cover image found in the first gallery.
The method for identifying a cover image according to claim 1, wherein:

The image of the predetermined area contained in the second gallery also has a first label of length and width;

The acquisition of each block image of the cover image to be identified includes:

Obtain the block position of each block of the cover image to be recognized;

Using the length and width included in the first tag, intercept each block of the cover image to be recognized to generate a block image.
The method for identifying a cover image according to claim 1, wherein:

The predetermined area image contained in the second gallery also has a second label, and the second label includes the first image characteristic information and the first text characteristic information of the predetermined area image;

The method further includes: when acquiring each block image of the to-be-recognized cover image, further acquiring second image feature information and second character feature information of the block image;

Searching for a predetermined area image identical to the block image in the second gallery, and identifying the to-be-recognized cover image as a sample cover image corresponding to the predetermined area image searched in the second gallery ,include:

Searching in the second gallery for a predetermined area image represented by a second label that is the same as the first image characteristic information and the first text characteristic information of the block image;

The cover image to be recognized is identified as a sample cover image corresponding to a predetermined area image found in the second library.
The method for identifying a cover image according to claim 1, wherein:

The method also includes: pre-establishing a second gallery;

The method for establishing the second gallery includes:

Obtain the first gallery, and generate a blank gallery;

Screening sample cover images with the same relevance in the first gallery;

Take a screenshot of the predetermined area on the screened sample cover image to obtain an image of the predetermined area;

Acquiring the length and width of the predetermined area image, and using the length and width to generate a first label of the predetermined area image;

extracting first image feature information and first text feature information of the predetermined area image, and using the first image feature information and the first text feature information to generate a second label of the predetermined area image;

distributing all the predetermined area images with the first label and the second label in the predetermined area of the blank gallery to generate a second gallery, in each predetermined area of the second gallery, between the sample cover images in each predetermined area are the same, and the correlations between sample cover images are different between different predetermined regions.
The method for identifying a cover image according to claim 1, wherein:

The method further includes: generating a pre-search result for the sample cover image searched in the first gallery in advance, and marking the pre-search result that the sample cover image searched in the first gallery is related to other sample cover images;

The judging whether the sample cover image searched in the first gallery is related to other sample cover images includes:

Acquire and search for a sample cover image matching the cover image to be identified in the first gallery, and obtain the same search result as the pre-search result;

identifying whether there is a mark in the pre-search result corresponding to the search result;

If so, it is determined that the sample cover image found in the first gallery is related to other sample cover images.
The method for identifying a cover image according to claim 1, wherein:

The cover image to be identified is the cover image of a picture book;

The method for obtaining the first gallery and the second gallery includes:

Read local storage, or/and, read cloud storage, and acquire the first gallery and the second gallery.
The method for identifying a cover image according to claim 1, wherein:

Both the predetermined area image and the block image include at least the name and serial number of the corresponding sample cover image.
A recognition device for a cover image, characterized in that it includes:

The data acquisition module is used to acquire the cover image to be identified, the first gallery and the second gallery, the first gallery of the cover image to be identified contains all sample cover images that have been included, and the second gallery contains at least one set of sample cover images The sample cover images corresponding to each group of predetermined area images have a correlation, and the sample cover images with the correlation include a series of sample cover images, or include a group of multilingual A sample cover image for ;

A first search module, configured to search for a sample cover image matching the cover image to be identified in the first gallery;

A judging module, configured to judge whether the sample cover image searched in the first gallery is related to other sample cover images;

The second search module obtains each block image of the cover image to be identified, searches for an image of a predetermined area identical to the block image in the second gallery, and identifies the cover image to be identified as The sample cover image corresponding to the image in the predetermined area searched out of the second gallery;

An output module, configured to identify the cover image to be recognized as the sample cover image found in the first library if the judgment of the judging module is negative.
A computer-readable storage medium storing a computer program, when the computer program is executed by a processor, the processor is made to perform the steps of the method according to any one of claims 1 to 7.
An image recognition device, comprising a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the process described in any one of claims 1 to 7. method steps.