CN110737792A - Exercise searching method, exercise searching device, exercise searching equipment and storage medium - Google Patents
Exercise searching method, exercise searching device, exercise searching equipment and storage medium Download PDFInfo
- Publication number
- CN110737792A CN110737792A CN201911004810.3A CN201911004810A CN110737792A CN 110737792 A CN110737792 A CN 110737792A CN 201911004810 A CN201911004810 A CN 201911004810A CN 110737792 A CN110737792 A CN 110737792A
- Authority
- CN
- China
- Prior art keywords
- picture
- sub
- searching
- processed
- vector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/40—Document-oriented image-based pattern recognition
- G06V30/41—Analysis of document content
- G06V30/413—Classification of content, e.g. text, photographs or tables
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30176—Document
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Library & Information Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the application discloses a problem searching method, a device, equipment and a storage medium, which relate to the technical field of artificial intelligence and comprise the steps of obtaining a picture to be processed, cutting the picture to be processed to obtain a sub-picture, confirming the th vector code of the sub-picture, searching an effective problem picture in a problem searching engine based on the th vector code to be used as a problem searching result of the picture to be processed, and solving the technical problem that problems containing pictures cannot be accurately searched in a problem bank in the prior art by adopting the scheme.
Description
Technical Field
The embodiment of the application relates to the technical field of artificial intelligence, in particular to a method, a device, equipment and a storage medium for searching for exercises.
Background
Generally, a user can shoot a picture with exercises through the family education machine, then the family education machine identifies characters in the picture, searches exercises which are the same as or similar to the shot exercises in an exercise library based on the characters, and returns the exercises to the user.
In summary, how to accurately search the problem including the picture becomes a problem which needs to be solved urgently.
Disclosure of Invention
The application provides exercises searching method, device and equipment and a storage medium, which are used for solving the technical problem that exercises containing pictures cannot be accurately searched in an exercise library in the prior art.
, the embodiment of the present application provides exercise searching methods, including:
acquiring a picture to be processed;
cutting the picture to be processed to obtain a sub-picture, wherein the sub-picture comprises exercise content;
confirming th vector encoding of the sub-picture;
and searching a valid problem picture in a problem search engine based on the th vector code as a problem search result of the picture to be processed.
, the problem content comprises at least sub-content pictures;
the cropping the picture to be processed to obtain a sub-picture comprises:
acquiring a picture frame and a color of each sub-content picture in the picture to be processed;
identifying all sub-content pictures within the same horizontal line range and/or the same vertical line range according to the picture borders and colors;
and clipping the picture to be processed based on all the sub-content pictures to obtain a sub-picture.
Further , the th vector encoding is a 128 bit code encoding.
, the finding a valid problem picture in a problem search engine as a problem search result of the picture to be processed based on the vector encoding includes:
inputting the th vector code into a problem search engine for searching, wherein the problem search engine comprises a plurality of second vector codes, and each second vector code corresponds to picture IDs;
obtaining a plurality of picture IDs and corresponding picture confidence levels according to output results of the problem search engine, wherein each picture ID corresponds to picture confidence levels, and the picture confidence levels are determined through the vector codes and the second vector codes corresponding to the picture IDs;
and determining a picture ID corresponding to an effective problem picture in the obtained plurality of picture IDs according to each picture confidence degree to serve as a problem search result.
Further , the determining, according to each of the picture confidences, a picture ID corresponding to a valid problem picture among the obtained plurality of picture IDs as a problem search result includes:
and judging the confidence of each image, and searching a set number of image IDs in the output plurality of image IDs according to the judgment result to be used as a problem search result.
, further comprising:
acquiring a picture exercise set, wherein the picture exercise set comprises a plurality of pictures, and each picture corresponds to picture exercises;
clipping each picture in the picture exercise set;
determining a second vector code for each of the cropped pictures;
and storing the second vector code into the problem search engine to obtain the problem search engine.
Further , the determining a second vector code for each of the cropped pictures includes:
determining second vector codes of each cut picture in batch by utilizing thread pool
In a second aspect, an embodiment of the present application further provides exercise question searching apparatuses, including:
the image acquisition module is used for acquiring an image to be processed;
the picture cutting module is used for cutting the picture to be processed to obtain a sub-picture, and the sub-picture comprises exercise content;
an encoding confirmation module for confirming the th vector encoding of the sub-picture;
and the problem searching module is used for searching a valid problem picture in a problem searching engine based on the th vector code as a problem searching result of the picture to be processed.
In a third aspect, an embodiment of the present application further provides exercise question searching apparatuses, including:
or more processors;
a memory for storing or more programs;
when the or more programs are executed by the or more processors, the or more processors implement the problem search method of aspect .
In a fourth aspect, embodiments of the present application further provide storage media containing computer-executable instructions that, when executed by a computer processor, perform the problem search method of aspect .
According to the problem searching method, the device, the equipment and the storage medium, the to-be-processed picture is obtained and cut to obtain the sub-picture containing the problem content, then the th vector code of the sub-picture is determined, and the technical scheme that the effective problem picture is searched in the problem searching engine to serve as the problem searching result of the to-be-processed picture based on the th vector code is adopted.
Drawings
FIG. 1 is a flowchart of the problem search method provided in embodiment of the present application;
FIG. 2 is a flowchart of a method for searching for problems according to the second embodiment of the present application;
fig. 3 is a schematic structural diagram of an problem search device according to a third embodiment of the present application;
fig. 4 is a schematic structural diagram of problem search devices according to the fourth embodiment of the present application.
Detailed Description
The present application is described in further detail below at with reference to the drawings and examples, it being understood that the specific examples described herein are for purposes of illustration and not limitation, and it is to be understood that only some, if not all, of the structures described herein are shown in the drawings for purposes of illustration.
It is noted that herein, relational terms such as th and second are used solely to distinguish entities or operations or objects from another entities or operations or objects, without necessarily requiring or implying any such actual relationship or order prior to such entities or operations or objects, " th" and "second" for vector code and second vector code are used to distinguish two different vector codes.
Example
Fig. 1 is a flowchart of problem searching methods provided in of the present application, where the problem searching methods provided in the embodiments can be executed by a problem searching apparatus, which can be implemented in software and/or hardware and integrated into a problem searching device, where the problem searching device can be an intelligent device with data processing and analyzing capabilities, such as a family education machine, a computer, a mobile phone, and the like.
Specifically, referring to fig. 1, the problem search method specifically includes:
and step 110, acquiring a picture to be processed.
In particular, the picture to be processed is a picture containing problems for searching for relevant problems, the picture to be processed comprises at least problems, in the embodiment, problems are described as an example, the problem can be a problem of a pure character type or a problem containing a picture type, wherein when the problem is a problem containing a picture type, the problem can contain pictures or a plurality of pictures, in the embodiment, the pictures contained in the problem are marked as subcontent pictures, namely, the picture to be processed comprises at least subcontent pictures, for example, the problem contains a problem stem part and an option part, the problem stem part is divided into apples larger than 50 which are painted black, the option part contains 6 apple pictures, and each apple contains sets of numbers, in this case, the problem can be determined as a picture type problem, and each apple corresponds to pictures, namely, 6 subcontent pictures.
, the user can take the picture to be processed by the problem searching device, at this time, the problem searching device includes at least cameras, or the problem searching device can communicate with other devices (such as mobile phone, computer, etc.) by data, at this time, the problem searching device can receive the picture to be processed sent by other devices.
Optionally, when the to-be-processed picture only includes exercises, after the to-be-processed picture is obtained, it may be determined whether the to-be-processed picture only includes exercises, if so, a subsequent operation is performed, otherwise, the user is prompted to cut the to-be-processed picture or to re-obtain the to-be-processed picture so that the to-be-processed picture only includes exercises.
And step 120, cutting the picture to be processed to obtain a sub-picture, wherein the sub-picture comprises exercise contents.
In an example, in order to ensure the accuracy of the result of the problem search, the picture to be processed is cut to eliminate invalid regions in the picture to be processed, wherein the invalid regions can be understood as regions with invalid characteristics, such as regions outside the contents of the problem, in step , after the picture to be processed is cut, the obtained picture is marked as a sub-picture, and the sub-picture is pictures with more accurate included characteristics relative to the picture to be processed.
Specifically, the means used when the picture to be processed is cropped may be set according to the actual situation, for example, by using an image recognition algorithm, to determine the horizontal line range and/or the vertical line range of each sub-content picture included in the problem, then determine the sub-content pictures in the same horizontal line range and/or the same vertical line range according to the horizontal line range and/or the vertical line range of each sub-content picture, and crop the picture to be processed, so that only the sub-content pictures in the same horizontal line range and/or the same vertical line range are included in the cropped sub-pictures.
And step 130, confirming th vector encoding of the sub-picture.
Specifically, after obtaining the sub-picture, encoding the sub-picture to obtain the feature vector corresponding to the sub-picture, in the embodiment, the obtained feature vector is recorded as th vector encoding, it can be understood that th vector encoding is code encoding with a set number of bits, in the embodiment, th vector encoding is exemplified as 128-bit code encoding, that is, after encoding the sub-picture, the sub-picture can be compressed to generate pictures with 128-bit code.
In particular, the problem search engine can be understood as a problem library which can be configured on the problem search device side or on a background server, when the problem search engine is configured on the background server, the problem search device can search in the problem search engine by accessing the background server, step , in the embodiment, the problem search engine can search for problems with pictures, wherein the embodiment of the establishment mode of the problem search engine is not limited, each problem search engine can correspond to problem sets, for example, the problem set corresponding to the problem search engine is a problem with pictures contained in a three-year math book of primary schools, or a problem with pictures contained in all teaching materials of primary schools, optionally, data search can be performed by a background technician and a problem set can be established, step , since the problem set contains problems with pictures in the embodiment, the problem set can also be recorded as a typical problem set, each problem set in the interior can be encoded with a vector , and each problem set can be obtained by encoding a vector of each problem set after the vector of the problem search engine is encoded with a vector 46id, and the vector ID of each problem set is encoded.
, after vector codes are obtained, vector codes are used as input of the problem search engine, then the problem search engine can search in the existing vector codes according to vector codes, wherein in the searching process, the confidence degrees of vector codes and all vector codes in the problem search engine are calculated, then a set number of vector codes with the highest confidence degrees are selected as the searching results of the problem search engine, then the problem IDs corresponding to the searched vector codes are obtained as the problem searching results, and the problem searching results are displayed for the user to select.
It can be understood that, in practical application, the method can also be adopted to search the to-be-processed picture containing the word exercises, and at this time, only the word content needs to be reserved when the to-be-processed picture is cut.
The method comprises the steps of obtaining a picture to be processed, cutting the picture to be processed to obtain a sub-picture containing problem content, determining th vector coding of the sub-picture, searching an effective problem picture in a problem search engine as a technical scheme of a problem search result of the picture to be processed based on th vector coding, cutting the picture to be processed to reduce an invalid region of the picture to be processed and ensure search accuracy, and searching through the coded picture by adopting a picture coding mode to solve the technical problem that the problem containing the picture cannot be accurately searched, so that the artificial intelligence search accuracy is improved, and the use experience of a user is improved.
Example two
Fig. 2 is a flowchart of problem searching methods provided in the second embodiment of the present application, the problem searching method provided in the present embodiment is embodied on the basis of the above embodiments, and specifically, referring to fig. 2, the problem searching method provided in the present embodiment includes:
and step 210, acquiring a picture to be processed.
And step 220, acquiring the picture frame and the color of each sub-content picture in the picture to be processed.
Specifically, the problem in the picture to be processed is set to include at least pictures, that is, at least sub-content pictures, at this time, the problem content included in the sub-picture obtained based on clipping should include at least sub-content pictures.
The image processing method comprises the following steps of exemplarily, YOLOv3 is a neural network model under a YOLO series, and a single sub-content picture in a picture to be processed can be identified through the YOLOv3 network model.
And step 230, identifying all sub-content pictures in the same horizontal line range and/or the same vertical line range according to picture borders and colors.
In the embodiment, it is described as an example that all sub-content pictures in the same horizontal line range and the same vertical line range are confirmed according to picture borders and colors.
The method comprises the steps of obtaining horizontal lines and vertical lines of each sub-content picture in a picture to be processed according to the position, and determining all sub-content pictures in the same horizontal line range and the same vertical line range according to the horizontal lines and the vertical lines of each sub-content picture.
And step 240, clipping the picture to be processed based on all the sub-content pictures to obtain sub-pictures.
Specifically, the graph cropping model extracts all the determined sub-content pictures in the same horizontal line range and the same vertical line range from the pictures to be processed, and further obtains sub-pictures.
For example, the problem corresponding to the picture to be processed includes 6 apples, and the 6 apples are arranged in 2 rows and 3 columns, it can be understood that each apple can be understood as sub-content pictures, specifically, after the picture to be processed is subjected to image processing, the color and picture frame of each apple can be obtained, and further the position of each apple in the picture to be processed is determined.
Step 250, confirming th vector encoding of the sub-picture.
The th vector code is a 128-bit code, specifically, the th vector code can be determined by a graph search algorithm model, the graph search algorithm model can be understood as a model of an integrated image coding algorithm, and the specific embodiment of the creation rule is not limited.
Step 260, inputting the th vector code into a problem search engine for searching, wherein the problem search engine comprises a plurality of second vector codes, and each second vector code corresponds to picture IDs.
Step , each second vector code corresponds to pictures containing problems, each picture containing problems corresponds to picture IDs, and therefore, each second vector code can correspond to picture IDs, and the corresponding relation can be and imported into the problem search engine, or the corresponding relation can be stored in a database, and the database can also store the second vector codes and each picture containing problems.
And 270, obtaining a plurality of picture IDs and corresponding picture confidence degrees according to the output result of the problem search engine, wherein each picture ID corresponds to picture confidence degrees, and the picture confidence degrees are determined through vector codes and second vector codes corresponding to the picture IDs.
The problem search engine can calculate the confidence of each second vector code and th vector code and record the confidence as picture confidence, wherein the embodiment of the calculation mode of the picture confidence is not limited, generally indicates that the higher the picture confidence, the higher the possibility that the second vector code is the same as or similar to the th vector code.
Meanwhile, each problem in the picture problem set corresponding to the problem search engine has a corresponding picture ID, so that the second vector code corresponding to each problem also has a corresponding picture ID., when the problem search engine outputs the picture confidence of each second vector code, the picture ID of each second vector code can be obtained based on the corresponding relationship between the second vector code and the picture ID, at this time, each picture confidence corresponds to pictures ID., or the picture ID corresponding to each second vector code is synchronously recorded in the problem search engine, at this time, the picture confidence and the corresponding picture ID of each second vector code can be synchronously output in the problem search engine.
And step 280, determining a picture ID corresponding to an effective problem picture from the obtained multiple picture IDs according to the confidence of each picture as a problem search result.
Specifically, the problem search engine may output the image confidence of all second vector codes or may output the image confidence of partial second vector codes. In this case, the corresponding picture IDs may be arranged in descending order according to the confidence of each picture output from the problem search engine, and then the arranged picture IDs may be fed back to the user.
At step , in practical application, the number of picture IDs fed back to the user can be set, and then the picture IDs with the highest picture confidence are selected according to the set number and fed back to the user.
Illustratively, the image confidences output by the problem search engine are arranged in the order from high to low. And then, selecting the confidence degrees of the pictures with the set number from high to low, and acquiring the corresponding picture ID as a problem search result to return to the user. Wherein, the set number can be set according to the actual situation. For example, the set number is 2.
Optionally, a confidence threshold may be preset, and then, the confidence of each image output by the problem search engine is compared with the confidence threshold to perform confidence judgment. And acquiring the confidence degrees of all the pictures higher than the confidence degree threshold, and if the number of the picture confidence degrees higher than the confidence degree threshold is less than or equal to the set number, directly taking the picture ID corresponding to the picture confidence degree higher than the confidence degree threshold as a problem search result. If the number of the image confidence levels higher than the confidence level threshold is greater than the set number, the image IDs corresponding to the set number of image confidence levels with the highest image confidence level can be selected from the image confidence levels higher than the confidence level threshold as the problem search result.
The method comprises the steps of obtaining a picture to be processed, obtaining all sub-content pictures in the same horizontal line range and/or the same vertical line range by identifying picture frames and colors of all sub-content pictures in the picture to be processed, cutting all the sub-content pictures according to all the sub-content pictures to obtain the sub-pictures, improving the accuracy of picture cutting, particularly when all the sub-content pictures in the same horizontal line range and the same vertical line range are obtained, enabling the sub-pictures to be up to 99.5%, then coding the sub-pictures to generate 128-bit code pictures, searching the problems in a problem search engine according to the 128-bit code pictures, searching the problems containing the pictures, ensuring the search accuracy, particularly when the problems contain a plurality of pictures, improving the accuracy by more than 70%, and improving the problem search experience of a user.
On the basis of the above embodiment, before performing the problem search, a picture search engine needs to be constructed. Therefore, the present embodiment further includes:
step 290, a picture problem set is obtained, wherein the picture problem set comprises a plurality of pictures, and each picture corresponds to picture problems.
The picture problem set comprises problem sets with a certain number, wherein the picture problem set comprises a plurality of pictures, each picture corresponds to problems, each problem is a problem with a picture, it can be understood that the embodiment of the obtaining mode of each problem contained in the picture problem set is not limited, and each problem corresponds to picture IDs.
Step 2100, crop each picture in the picture problem set.
Specifically, each picture in the picture problem set is cut according to the cutting mode provided above, so as to obtain a sub-picture corresponding to each picture.
Step 2110, determining a second vector code of each cut picture.
It is understood that each picture corresponds to second vector codes, i.e., second vector codes per picture ID.
Considering that the number of the problems included in the picture problem set is large, if image coding is performed on each cut picture in sequence, the data processing amount is too large, and therefore, in the embodiment, the determining the second vector coding of each cut picture includes: and determining the second vector code of each cut picture in batch by using the thread pool.
The thread pool is multithread processing forms, tasks are added into a queue in the processing process, and then the tasks are automatically started after the threads are created.
Step 2120, storing the second vector code in a problem search engine to obtain a problem search engine.
Specifically, when the second vector code is stored in the problem search engine, the second vector code can be stored in batches by using the thread pool, so that the storage speed is improved.
In the above, the problem search engine is further constructed by constructing the second vector code, and then the problem including the picture is searched by the problem search engine, so that the problem including the picture can be accurately searched.
EXAMPLE III
Fig. 3 is a schematic structural diagram of problem search devices according to the third embodiment of the present invention, and referring to fig. 3, the problem search device according to the present embodiment includes a picture obtaining module 301, a picture cropping module 302, a code confirmation module 303, and a problem search module 304.
The image processing device comprises a picture acquisition module 301 for acquiring a picture to be processed, a picture cutting module 302 for cutting the picture to be processed to obtain a sub-picture, a coding confirmation module 303 for confirming th vector coding of the sub-picture, and a problem search module 304 for searching an effective problem picture in a problem search engine based on th vector coding as a problem search result of the picture to be processed.
The method comprises the steps of obtaining a picture to be processed, cutting the picture to be processed to obtain a sub-picture containing problem content, determining th vector coding of the sub-picture, searching an effective problem picture in a problem search engine as a technical scheme of a problem search result of the picture to be processed based on th vector coding, cutting the picture to be processed to reduce an invalid region of the picture to be processed and ensure search accuracy, and searching through the coded picture by adopting a picture coding mode to solve the technical problem that the problem containing the picture cannot be accurately searched, so that the artificial intelligence search accuracy is improved, and the use experience of a user is improved.
On the basis of the embodiment, the problem content comprises at least sub-content pictures, the picture cropping module 302 comprises a data acquisition unit for acquiring a picture frame and a color of each sub-content picture in the picture to be processed, a content confirmation unit for confirming all sub-content pictures in the same horizontal line range and/or the same vertical line range according to the picture frame and the color, and a content cropping unit for cropping the picture to be processed based on all the sub-content pictures to obtain sub-pictures.
On the basis of the above embodiment, the th vector code is a 128-bit code.
On the basis of the embodiment, the problem searching module 304 comprises a vector input unit, a confidence coefficient output unit and a result determining unit, wherein the vector input unit is used for inputting the th vector code into a problem searching engine for searching, the problem searching engine comprises a plurality of second vector codes, each second vector code corresponds to picture IDs, the confidence coefficient output unit is used for obtaining a plurality of picture IDs and corresponding picture confidence coefficients according to the output result of the problem searching engine, each picture ID corresponds to picture confidence coefficients, and the picture confidence coefficients are determined through the th vector code and the second vector code corresponding to the picture IDs, and the result determining unit is used for determining the picture ID corresponding to a valid problem picture as a problem searching result in the obtained plurality of picture IDs according to each picture confidence coefficient.
On the basis of the foregoing embodiment, the result determining unit is specifically configured to: and judging the confidence of each image, and searching a set number of image IDs in the output plurality of image IDs according to the judgment result to be used as a problem search result.
On the basis of the embodiment, the system further comprises a set acquisition module, a set cutting module, a code determination module and an engine construction module, wherein the set acquisition module is used for acquiring a picture exercise set, the picture exercise set comprises a plurality of pictures, each picture corresponds to picture exercises, the set cutting module is used for cutting each picture in the picture exercise set, the code determination module is used for determining second vector codes of each cut picture, and the engine construction module is used for storing the second vector codes into the exercise search engine to obtain the exercise search engine.
On the basis of the foregoing embodiment, the code determining module is specifically configured to: and determining the second vector code of each cut picture in batch by utilizing the thread pool.
The problem searching device provided by the embodiment is included in the problem searching equipment, can be used for executing the problem searching method provided by any embodiment, and has corresponding functions and beneficial effects.
Example four
Fig. 4 is a schematic structural diagram of problem searching apparatuses according to a fourth embodiment of the present invention, specifically, as shown in fig. 4, the problem searching apparatus includes a processor 40, a memory 41, an input device 42, an output device 43, and a communication device 44, the number of processors 40 in the problem searching apparatus may be or more, processors 40 in fig. 4 are taken as an example, the processor 40, the memory 41, the input device 42, the output device 43, and the communication device 44 in the problem searching apparatus may be connected by a bus or other means, and fig. 4 is taken as an example of connection by a bus.
Memory 4 is used to store software programs, computer-executable programs, and modules, such as program instructions/modules in the problem search method in the embodiments of the present application (e.g., picture taking module 301, picture cropping module 302, code confirmation module 303, and problem search module 304 in the problem search device), as computer-readable storage media, processor 40 executes various functional applications and data processing of the problem search device by running the software programs, instructions, and modules stored in memory 41, that is, implementing the problem search method provided by any of the embodiments described above.
The memory 41 may generally include a program storage area that may store an operating system, at least applications needed for functionality, and a data storage area that may store data created from use of the problem search device, etc. additionally, the memory 41 may include high speed random access memory, and may also include non-volatile memory, such as at least disk storage devices, flash memory devices, or other non-volatile solid state storage devices in examples, the memory 41 may further include memory remotely located from the processor 40 that may be connected to the problem search device via a network examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The input device 42 is operable to receive input numeric or character information and to generate key signal inputs relating to user settings and function controls of the problem-searching apparatus. The output device 43 may include a display screen, a speaker, etc. The communication means 44 is used for data communication with a background server or other devices.
The problem searching device comprises the problem searching device provided by the third embodiment, can be used for executing the problem searching method provided by any embodiment, and has corresponding functions and beneficial effects.
EXAMPLE five
Embodiments of the present application also provide storage media containing computer-executable instructions that, when executed by a computer processor, perform problem search methods, the methods comprising:
acquiring a picture to be processed;
cutting the picture to be processed to obtain a sub-picture, wherein the sub-picture comprises exercise content;
confirming th vector encoding of the sub-picture;
and searching a valid problem picture in a problem search engine based on the th vector code as a problem search result of the picture to be processed.
Of course, the storage media containing computer-executable instructions provided in the embodiments of the present application are not limited to the method operations described above, and may also perform related operations in the problem search method provided in any embodiments of the present application.
Based on the understanding that the technical solutions of the present application can be embodied in the form of software products, such as floppy disks, Read-Only memories (ROMs), Random Access Memories (RAMs), FLASH memories (flashes), hard disks, optical disks, etc., which are stored in a computer-readable storage medium, and include instructions for enabling computer devices (which may be personal computers, servers, or network devices, etc.) to execute the method for searching the problems described in the embodiments of the present application.
It should be noted that, in the embodiment of the problem search device, the included units and modules are only divided according to functional logic, but are not limited to the above division as long as the corresponding functions can be realized; in addition, specific names of the functional units are only used for distinguishing one functional unit from another, and are not used for limiting the protection scope of the application.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present application and the technical principles employed. It will be understood by those skilled in the art that the present application is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the application. Therefore, although the present application has been described in more detail with reference to the above embodiments, the present application is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present application, and the scope of the present application is determined by the scope of the appended claims.
Claims (10)
1, exercise searching method, comprising:
acquiring a picture to be processed;
cutting the picture to be processed to obtain a sub-picture, wherein the sub-picture comprises exercise content;
confirming th vector encoding of the sub-picture;
and searching a valid problem picture in a problem search engine based on the th vector code as a problem search result of the picture to be processed.
2. The problem search method of claim 1, wherein the problem content comprises at least sub-content pictures;
the cropping the picture to be processed to obtain a sub-picture comprises:
acquiring a picture frame and a color of each sub-content picture in the picture to be processed;
identifying all sub-content pictures within the same horizontal line range and/or the same vertical line range according to the picture borders and colors;
and clipping the picture to be processed based on all the sub-content pictures to obtain a sub-picture.
3. The problem search method of claim 1, wherein said th vector code is a 128-bit code.
4. The problem search method of claim 1, wherein said finding a valid problem picture in a problem search engine as a result of the problem search of the picture to be processed based on the th vector encoding comprises:
inputting the th vector code into a problem search engine for searching, wherein the problem search engine comprises a plurality of second vector codes, and each second vector code corresponds to picture IDs;
obtaining a plurality of picture IDs and corresponding picture confidence levels according to output results of the problem search engine, wherein each picture ID corresponds to picture confidence levels, and the picture confidence levels are determined through the vector codes and the second vector codes corresponding to the picture IDs;
and determining a picture ID corresponding to an effective problem picture in the obtained plurality of picture IDs according to each picture confidence degree to serve as a problem search result.
5. The problem search method of claim 4, wherein the determining, according to each of the image confidences, an image ID corresponding to a valid problem image among the obtained plurality of image IDs as a problem search result comprises:
and judging the confidence of each image, and searching a set number of image IDs in the output plurality of image IDs according to the judgment result to be used as a problem search result.
6. The problem search method according to claim 4, further comprising:
acquiring a picture exercise set, wherein the picture exercise set comprises a plurality of pictures, and each picture corresponds to picture exercises;
clipping each picture in the picture exercise set;
determining a second vector code for each of the cropped pictures;
and storing the second vector code into the problem search engine to obtain the problem search engine.
7. The problem search method of claim 6, wherein said determining a second vector encoding for each of said cropped pictures comprises:
and determining the second vector code of each cut picture in batch by utilizing the thread pool.
The problem search device of kinds, characterized by comprising:
the image acquisition module is used for acquiring an image to be processed;
the picture cutting module is used for cutting the picture to be processed to obtain a sub-picture, and the sub-picture comprises exercise content;
an encoding confirmation module for confirming the th vector encoding of the sub-picture;
and the problem searching module is used for searching a valid problem picture in a problem searching engine based on the th vector code as a problem searching result of the picture to be processed.
An apparatus for searching for problems of the type 9, , comprising:
or more processors;
a memory for storing or more programs;
when executed by the or more processors, the or more programs cause the or more processors to implement the problem search method of any of claims 1-7 to .
A storage medium containing computer-executable instructions, wherein the computer-executable instructions, when executed by a computer processor, are for performing the problem search method of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911004810.3A CN110737792A (en) | 2019-10-22 | 2019-10-22 | Exercise searching method, exercise searching device, exercise searching equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911004810.3A CN110737792A (en) | 2019-10-22 | 2019-10-22 | Exercise searching method, exercise searching device, exercise searching equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110737792A true CN110737792A (en) | 2020-01-31 |
Family
ID=69270763
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911004810.3A Pending CN110737792A (en) | 2019-10-22 | 2019-10-22 | Exercise searching method, exercise searching device, exercise searching equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110737792A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107729491A (en) * | 2017-10-18 | 2018-02-23 | 广东小天才科技有限公司 | Improve the method, apparatus and equipment of the accuracy rate of topic answer search |
| CN108228757A (en) * | 2017-12-21 | 2018-06-29 | 北京市商汤科技开发有限公司 | Image search method and device, electronic equipment, storage medium, program |
| CN109492644A (en) * | 2018-10-16 | 2019-03-19 | 深圳壹账通智能科技有限公司 | A kind of matching and recognition method and terminal device of exercise image |
| CN110085068A (en) * | 2019-04-22 | 2019-08-02 | 广东小天才科技有限公司 | A kind of study coach method and device based on image recognition |
| CN110263199A (en) * | 2019-06-21 | 2019-09-20 | 君库(上海)信息科技有限公司 | It is a kind of based on the cartographical sketching of deep learning to scheme to search drawing method |
-
2019
- 2019-10-22 CN CN201911004810.3A patent/CN110737792A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107729491A (en) * | 2017-10-18 | 2018-02-23 | 广东小天才科技有限公司 | Improve the method, apparatus and equipment of the accuracy rate of topic answer search |
| CN108228757A (en) * | 2017-12-21 | 2018-06-29 | 北京市商汤科技开发有限公司 | Image search method and device, electronic equipment, storage medium, program |
| CN109492644A (en) * | 2018-10-16 | 2019-03-19 | 深圳壹账通智能科技有限公司 | A kind of matching and recognition method and terminal device of exercise image |
| CN110085068A (en) * | 2019-04-22 | 2019-08-02 | 广东小天才科技有限公司 | A kind of study coach method and device based on image recognition |
| CN110263199A (en) * | 2019-06-21 | 2019-09-20 | 君库(上海)信息科技有限公司 | It is a kind of based on the cartographical sketching of deep learning to scheme to search drawing method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021017998A1 (en) | Method and system for positioning text position, and method and system for training model | |
| US11450027B2 (en) | Method and electronic device for processing videos | |
| US9222783B2 (en) | Location estimation using image analysis | |
| US20080101723A1 (en) | Coverage-Based Image Relevance Ranking | |
| CN110347866B (en) | Information processing method, information processing device, storage medium and electronic equipment | |
| CN112199582B (en) | Content recommendation method, device, equipment and medium | |
| CN112818995B (en) | Image classification method, device, electronic equipment and storage medium | |
| CN111783712A (en) | Video processing method, device, equipment and medium | |
| CN110443170B (en) | Human body key point determining method and device and electronic equipment | |
| CN111881740B (en) | Face recognition method, device, electronic equipment and medium | |
| CN107729491B (en) | Method, device and equipment for improving accuracy rate of question answer search | |
| CN113157956B (en) | Picture searching method, system, mobile terminal and storage medium | |
| CN110895602B (en) | Identity authentication method and device, electronic equipment and storage medium | |
| CN114821613A (en) | Extraction method and system of table information in PDF | |
| CN112052352B (en) | Video ordering method, device, server and storage medium | |
| CN111539390A (en) | Small target image identification method, equipment and system based on Yolov3 | |
| CN110737792A (en) | Exercise searching method, exercise searching device, exercise searching equipment and storage medium | |
| CN111008295A (en) | Page retrieval method and device, electronic equipment and storage medium | |
| CN110895691A (en) | Image processing method and device and electronic equipment | |
| CN115272923A (en) | Intelligent identification method and system based on big data platform | |
| CN112464894B (en) | Interaction method and device and computer equipment | |
| CN116433939B (en) | Sample image generation method, training method, recognition method and device | |
| CN111918137B (en) | Push method and device based on video characteristics, storage medium and terminal | |
| CN111382306B (en) | Method and device for inquiring video frame | |
| CN117315771A (en) | Mobile phone behavior recognition method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200131 |
|
| RJ01 | Rejection of invention patent application after publication |