CN111553334A - Questionnaire image recognition method, electronic device, and storage medium - Google Patents
Questionnaire image recognition method, electronic device, and storage medium Download PDFInfo
- Publication number
- CN111553334A CN111553334A CN202010316545.9A CN202010316545A CN111553334A CN 111553334 A CN111553334 A CN 111553334A CN 202010316545 A CN202010316545 A CN 202010316545A CN 111553334 A CN111553334 A CN 111553334A
- Authority
- CN
- China
- Prior art keywords
- questionnaire
- image
- area
- identification
- template image
- 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.)
- Withdrawn
Links
Images
Classifications
-
- 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
- G06V30/10—Character recognition
- G06V30/22—Character recognition characterised by the type of writing
- G06V30/224—Character recognition characterised by the type of writing of printed characters having additional code marks or containing code marks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
-
- 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
- G06V30/10—Character recognition
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to an image recognition technology and provides a questionnaire image recognition method, an electronic device and a storage medium. The method comprises the steps of performing binarization processing and preset mode processing on a questionnaire image to be recognized, matching and correcting the processed questionnaire image to be recognized and a questionnaire template image to obtain a questionnaire image to be recognized, wherein the questionnaire image to be recognized and the questionnaire template image have corresponding coordinate positions, cutting the questionnaire image to be recognized based on the coordinates of the questionnaire template image, inputting the obtained ticket identification area and page identification area into a pre-trained recognition model to obtain recognition results of the ticket identification and the page identification, segmenting a content filling area based on the recognition results to obtain a plurality of evaluation areas, calculating the pixel value of a fillable area of each evaluation area, and taking the fillable area with the minimum pixel value as the result of the evaluation area. The invention has low dependence on hardware environment and high identification precision on questionnaire images.
Description
Technical Field
The present invention relates to the field of image recognition technologies, and in particular, to a method for recognizing a questionnaire image, an electronic device, and a storage medium.
Background
At present, when an investigation institution carries out statistical analysis on questionnaires, the questionnaires are mostly required to be manually input into an analysis system, the manual input mode is low in efficiency, and mistakes are easy to make. Although some automatic questionnaire recognition schemes are available in the market, the existing recognition schemes are usually based on optical recognition schemes, and depend on specific positioning areas, such as color blocks, anchor points, and the like, if the positioning points are not completely scanned or are missing due to other reasons, positioning fails, and operations such as rectification and alignment cannot be performed on the images, so that key information of the questionnaire images cannot be recognized.
Disclosure of Invention
In view of the above, the present invention provides a questionnaire image recognition method, an electronic device, and a storage medium, which are intended to solve the problems of high hardware requirements and low recognition accuracy in recognition of a questionnaire image in the related art.
In order to achieve the above object, the present invention provides a questionnaire image recognition method, comprising:
a receiving step: receiving a request for identifying a questionnaire image sent by a user, and acquiring the questionnaire image to be identified in the request;
a pretreatment step: performing binarization processing on the questionnaire image to be identified and a preset questionnaire template image, and performing processing in a preset mode on the questionnaire image to be identified and the questionnaire template image after binarization processing;
matching: matching and correcting the processed questionnaire image to be recognized with the questionnaire template image to obtain a questionnaire image to be recognized with a corresponding coordinate position of the questionnaire template image, obtaining coordinates of a ticket identification area, a page identification area and a content filling area of the image to be recognized based on the coordinates of the questionnaire template image, and cutting the ticket identification area, the page identification area and the content filling area of the questionnaire image to be recognized; and
an identification step: inputting the ticket identification area and the page identification area into a pre-trained identification model to obtain identification results of the ticket identification and the page identification of the questionnaire image to be identified, dividing the content filling area based on the identification results and the coordinates of the questionnaire template image to obtain a plurality of evaluation areas, calculating the pixel value of the fillable area of each evaluation area, and taking the fillable area with the minimum pixel value as the result of the evaluation area and feeding back the result to the user.
Preferably, the matching step includes:
analyzing the feature points of the questionnaire image to be identified and the questionnaire template image according to an SIFT algorithm, and respectively calculating the Euclidean distance between each feature point in the questionnaire template image and each feature point in the questionnaire image to be identified;
searching out two feature points with the minimum distance and the minimum distance in the questionnaire template image from each feature point in the questionnaire template image, and if the feature point with the minimum distance and the feature point with the minimum distance meet preset conditions, taking the feature point with the minimum distance and the corresponding feature point in the questionnaire template image as a pair of matched feature points;
and taking the questionnaire template image with the logarithm of the matching feature points larger than a first preset value as the image matched with the questionnaire image to be identified.
Preferably, the matching step includes:
respectively obtaining coordinates of matching feature points of the matched questionnaire image to be identified and the questionnaire template image, calculating to obtain a homography matrix for converting the image to be matched into the template image based on the coordinates of the matching feature points, and carrying out perspective transformation on the image to be matched based on the homography matrix to obtain the questionnaire image to be identified with the questionnaire template image having a corresponding coordinate position.
Preferably, the identifying step includes:
the recognition model comprises a page number recognition model and a ticket identification model, the page number identification area is input into the page number recognition model to obtain a recognition result of the page number identification of the questionnaire image to be recognized, and the ticket identification area is input into the ticket identification model to obtain a recognition result of the ticket identification of the questionnaire image to be recognized.
Preferably, the identifying step further comprises:
judging whether the plurality of evaluation areas have filling missing areas or not by using a first preset judgment rule, judging whether the plurality of evaluation areas have filling areas or not by using a second preset judgment rule, and if the filling missing areas or the filling areas exist, adding the filling missing areas or the filling areas to the result of the evaluation area.
To achieve the above object, the present invention also provides an electronic device, including: the questionnaire image recognition system comprises a memory and a processor, wherein the memory stores a questionnaire image recognition program, and the questionnaire image recognition program is executed by the processor to realize the following steps:
a receiving step: receiving a request for identifying a questionnaire image sent by a user, and acquiring the questionnaire image to be identified in the request;
a pretreatment step: performing binarization processing on the questionnaire image to be identified and a preset questionnaire template image, and performing processing in a preset mode on the questionnaire image to be identified and the questionnaire template image after binarization processing;
matching: matching and correcting the processed questionnaire image to be recognized with the questionnaire template image to obtain a questionnaire image to be recognized with a corresponding coordinate position of the questionnaire template image, obtaining coordinates of a ticket identification area, a page identification area and a content filling area of the image to be recognized based on the coordinates of the questionnaire template image, and cutting the ticket identification area, the page identification area and the content filling area of the questionnaire image to be recognized; and
an identification step: inputting the ticket identification area and the page identification area into a pre-trained identification model to obtain identification results of the ticket identification and the page identification of the questionnaire image to be identified, dividing the content filling area based on the identification results and the coordinates of the questionnaire template image to obtain a plurality of evaluation areas, calculating the pixel value of the fillable area of each evaluation area, and taking the fillable area with the minimum pixel value as the result of the evaluation area and feeding back the result to the user.
Preferably, the matching step includes:
analyzing the feature points of the questionnaire image to be identified and the questionnaire template image according to an SIFT algorithm, and respectively calculating the Euclidean distance between each feature point in the questionnaire template image and each feature point in the questionnaire image to be identified;
searching out two feature points with the minimum distance and the minimum distance in the questionnaire template image from each feature point in the questionnaire template image, and if the feature point with the minimum distance and the feature point with the minimum distance meet preset conditions, taking the feature point with the minimum distance and the corresponding feature point in the questionnaire template image as a pair of matched feature points;
and taking the questionnaire template image with the logarithm of the matching feature points larger than a first preset value as the image matched with the questionnaire image to be identified.
Preferably, the matching step includes:
respectively obtaining coordinates of matching feature points of the matched questionnaire image to be identified and the questionnaire template image, calculating to obtain a homography matrix for converting the image to be matched into the template image based on the coordinates of the matching feature points, and carrying out perspective transformation on the image to be matched based on the homography matrix to obtain the questionnaire image to be identified with the questionnaire template image having a corresponding coordinate position.
Preferably, the identifying step includes:
the recognition model comprises a page number recognition model and a ticket identification model, the page number identification area is input into the page number recognition model to obtain a recognition result of the page number identification of the questionnaire image to be recognized, and the ticket identification area is input into the ticket identification model to obtain a recognition result of the ticket identification of the questionnaire image to be recognized.
To achieve the above object, the present invention further provides a computer-readable storage medium, which includes a questionnaire image recognition program, and when the questionnaire image recognition program is executed by a processor, the computer-readable storage medium realizes any of the steps in the questionnaire image recognition method as described above.
The questionnaire image recognition method, the electronic device and the storage medium have the advantages of independence on hardware environment, high robustness, high recognition accuracy and the like, the questionnaire image to be recognized is corrected by taking the questionnaire template as a reference, the evaluation area of the questionnaire image to be recognized is analyzed through pixel analysis, the single-selection, multi-selection or omission area in the questionnaire image is recognized, and finally the ticket type and page number area in the questionnaire image is recognized through a deep learning technology, so that the extraction of the questionnaire image information is completed.
Drawings
FIG. 1 is a diagram of an electronic device according to a preferred embodiment of the present invention;
FIG. 2 is a block diagram of a preferred embodiment of the questionnaire image recognition program of FIG. 1;
FIG. 3 is a flow chart of a preferred embodiment of the questionnaire image recognition method of the invention;
the implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, a schematic diagram of an electronic device 1 according to a preferred embodiment of the invention is shown.
The electronic device 1 includes but is not limited to: memory 11, processor 12, display 13, and network interface 14. The electronic device 1 is connected to a network through a network interface 14 to obtain raw data. The network may be a wireless or wired network such as an Intranet (Intranet), the Internet (Internet), a Global System for mobile communications (GSM), Wideband Code Division Multiple Access (WCDMA), a 4G network, a 5G network, Bluetooth (Bluetooth), Wi-Fi, or a communication network.
The memory 11 includes at least one type of readable storage medium including a flash memory, a hard disk, a multimedia card, a card type memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Read Only Memory (ROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), a Programmable Read Only Memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the storage 11 may be an internal storage unit of the electronic device 1, such as a hard disk or a memory of the electronic device 1. In other embodiments, the memory 11 may also be an external storage device of the electronic apparatus 1, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like provided in the electronic apparatus 1. Of course, the memory 11 may also comprise both an internal memory unit of the electronic apparatus 1 and an external memory device thereof. In this embodiment, the memory 11 is generally used for storing an operating system installed in the electronic device 1 and various types of application software, such as a program code of the questionnaire image recognition program 10. Further, the memory 11 may also be used to temporarily store various types of data that have been output or are to be output.
Processor 12 may be a Central Processing Unit (CPU), controller, microcontroller, microprocessor, or other data Processing chip in some embodiments. The processor 12 is generally used for controlling the overall operation of the electronic device 1, such as performing data interaction or communication related control and processing. In this embodiment, the processor 12 is configured to run the program code or the processing data stored in the memory 11, for example, the program code of the questionnaire image recognition program 10.
The display 13 may be referred to as a display screen or display unit. In some embodiments, the display 13 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an Organic Light-emitting diode (OLED) touch screen, or the like. The display 13 is used for displaying information processed in the electronic device 1 and for displaying a visual work interface, for example, results of data statistics.
The network interface 14 may optionally comprise a standard wired interface, a wireless interface (e.g. WI-FI interface), the network interface 14 typically being used for establishing a communication connection between the electronic apparatus 1 and other electronic devices.
Fig. 1 shows only electronic device 1 with components 11-14 and questionnaire image recognition program 10, but it should be understood that not all of the shown components are required to be implemented, and that more or fewer components may be implemented instead.
Optionally, the electronic device 1 may further comprise a user interface, the user interface may comprise a Display (Display), an input unit such as a Keyboard (Keyboard), and the optional user interface may further comprise a standard wired interface, a wireless interface. Alternatively, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an Organic Light-Emitting Diode (OLED) touch screen, or the like. The display, which may also be referred to as a display screen or display unit, is suitable for displaying information processed in the electronic apparatus 1 and for displaying a visualized user interface.
The electronic device 1 may further include a Radio Frequency (RF) circuit, a sensor, an audio circuit, and the like, which are not described in detail herein.
In the above embodiment, the processor 12, when executing the questionnaire image recognition program 10 stored in the memory 11, may implement the following steps:
a receiving step: receiving a request for identifying a questionnaire image sent by a user, and acquiring the questionnaire image to be identified in the request;
a pretreatment step: performing binarization processing on the questionnaire image to be identified and a preset questionnaire template image, and performing processing in a preset mode on the questionnaire image to be identified and the questionnaire template image after binarization processing;
matching: matching and correcting the processed questionnaire image to be recognized with the questionnaire template image to obtain a questionnaire image to be recognized with a corresponding coordinate position of the questionnaire template image, obtaining coordinates of a ticket identification area, a page identification area and a content filling area of the image to be recognized based on the coordinates of the questionnaire template image, and cutting the ticket identification area, the page identification area and the content filling area of the questionnaire image to be recognized; and
an identification step: inputting the ticket identification area and the page identification area into a pre-trained identification model to obtain identification results of the ticket identification and the page identification of the questionnaire image to be identified, dividing the content filling area based on the identification results and the coordinates of the questionnaire template image to obtain a plurality of evaluation areas, calculating the pixel value of the fillable area of each evaluation area, and taking the fillable area with the minimum pixel value as the result of the evaluation area and feeding back the result to the user.
The storage device may be the memory 11 of the electronic apparatus 1, or may be another storage device communicatively connected to the electronic apparatus 1.
For a detailed description of the above steps, please refer to the following description of fig. 2 regarding a program block diagram of an embodiment of the questionnaire image recognition program 10 and fig. 3 regarding a flowchart of an embodiment of the questionnaire image recognition method.
In other embodiments, the questionnaire image recognition program 10 can be divided into a plurality of modules that are stored in the memory 12 and executed by the processor 13 to accomplish the present invention. The modules referred to herein are referred to as a series of computer program instruction segments capable of performing specified functions.
Referring to FIG. 2, a block diagram of an embodiment of the questionnaire image recognition program 10 of FIG. 1 is shown. In this embodiment, the questionnaire image recognition program 10 can be divided into: a receiving module 110, a preprocessing module 120, a matching module 130, and an identifying module 140.
The receiving module 110 is configured to receive a request for identifying a questionnaire image sent by a user, and acquire a questionnaire image to be identified in the request.
In this embodiment, the questionnaire image to be identified, which is captured by the user through the scanner or the mobile phone, can be received, and the system provided by the invention has strong robustness and low requirements on illumination and capturing angles, so that the mode and environment for the user to capture the questionnaire image to be identified are not limited.
The questionnaire image to be identified can comprise a primary table or a secondary table, and the structure of the primary table is as follows: the single-class individual multi-evaluation item can be scored in a single table, and the structure of a secondary table is as follows: multiple evaluation items for multiple classes of individuals can be scored in a single table. In the questionnaire tabulation process of the embodiment, a user can input items to be evaluated and individuals to be evaluated according to requirements in a questionnaire style, and ticket identification, for example, named by English capital letters A-Z, can be added in the upper right corner of a table in order to meet the statistical requirements of the same user on different groups to be evaluated in a classified manner. The table can be stored as a picture or printed for use after being automatically generated, and a JSON format input evaluation information file can be stored for subsequent identification statistics and next reuse. The storage mode can also facilitate the situation that tabulation and recognition statistics work are processed by different people respectively, and a user can carry out subsequent function operation only by opening the stored file in the JSON format by using the system.
The preprocessing module 120 is configured to perform binarization processing on the questionnaire image to be identified and a preset questionnaire template image, and then perform processing in a preset manner on the questionnaire image to be identified and the questionnaire template image after the binarization processing.
In this embodiment, the size of the questionnaire image to be recognized is adjusted to an image with the same size as the preset questionnaire template image, adaptive binarization processing is performed on the image to be recognized and the questionnaire template image by using a cv2.thresh _ OTSU method, and the image is converted into pixel values of 0 and 255, so that interferences of different illumination conditions are removed, and then preset mode processing is performed on the questionnaire image to be recognized and the questionnaire template image after binarization processing, wherein the preset mode processing includes performing corrosion processing on a check image adopting (5, 5), and interferences such as noise points and the like can be removed. The preprocessing operation enables the subsequent image matching to be more robust and not easily affected by illumination and printing ink depth.
The matching module 130 is configured to match and correct the processed questionnaire image to be recognized with the questionnaire template image to obtain a questionnaire image to be recognized having a corresponding coordinate position with the questionnaire template image, obtain coordinates of the ticket identification area, the page identification area, and the content filling area of the image to be recognized based on the coordinates of the questionnaire template image, and cut the ticket identification area, the page identification area, and the content filling area of the questionnaire image to be recognized.
In the embodiment, based on the consideration of both stability and speed, the image matching is performed by using the SIFT feature matching method. Specifically, feature points of a questionnaire image to be recognized and a questionnaire template image are analyzed according to an SIFT algorithm, Euclidean distances between each feature point in the questionnaire template image and each feature point in the questionnaire image to be recognized are respectively calculated, two feature points with the minimum distance and the minimum distance in the questionnaire template image and the questionnaire image to be recognized are found out, if the feature point with the minimum distance and the feature point with the minimum distance meet preset conditions, the feature point with the minimum distance and the corresponding feature point in the questionnaire template image are used as a pair of matching feature points, and the questionnaire template image with the logarithm of the matching feature points larger than a first preset value is used as the image matched with the questionnaire image to be recognized. Each feature point in the questionnaire template image and two feature points with the minimum distance and the second minimum distance in the questionnaire image to be identified can be searched by using a knmatch method in cv2.BFMatcher, and the preset conditions are as follows: a value obtained by dividing the distance value of the feature point having the smallest distance by the distance value of the feature point having the next smallest distance is smaller than a second preset value, and the second preset value may be 0.65. If the logarithm of the matching feature points is greater than the first preset value, it indicates that the two image contents include the same feature object, and the first preset value of this embodiment may be set to 4.
After the two images are determined to be matched, the questionnaire image to be recognized and the questionnaire template image are corrected, the coordinates of each feature point of the questionnaire image to be recognized and the questionnaire template image which are successfully matched are respectively obtained, a homography matrix of the template image converted from the image to be matched is obtained through calculation based on the coordinates of each matched feature point, perspective transformation is carried out on the image to be matched based on the homography matrix, and the questionnaire image to be recognized with the questionnaire template image in the corresponding coordinate position is obtained. The method of cv2.RANSAC can be adopted when calculating the homography matrix. The method comprises the steps of obtaining position coordinates of a ticket type identification area, a page number identification area and a content filling area of a questionnaire image to be identified based on coordinates of a ticket type, a page number and a content area of the questionnaire template image, and cutting the ticket type identification area, the page number identification area and the content filling area of the questionnaire image to be identified.
The recognition module 140 is configured to input the ticket identification area and the page identification area into a pre-trained recognition model, obtain recognition results of the ticket identification and the page identification of the questionnaire image to be recognized, divide the content filling area based on the recognition results and the coordinates of the questionnaire template image to obtain a plurality of evaluation areas, calculate a pixel value of a fillable area of each evaluation area, and feed back the fillable area with the smallest pixel value as a result of the evaluation area to the user.
In this embodiment, the ticket identification area and the page identification area of the questionnaire image to be recognized are input into a recognition model trained in advance based on a convolutional recurrent neural network, so as to obtain the recognition results of the ticket identification and the page identification of the questionnaire image to be recognized. The table type is divided into 26 letters from A to Z, and the table page numbers are numbers of 1, 2, 3 and the like. The CNN module based on VGG16 in the CRNN is replaced by Resnet101, and the Shotcut structure in Resnet can effectively prevent gradient disappearance, so that character features with higher distinguishing degree can be extracted by adopting a deeper network structure, and the problem of identification errors (such as I and L) of similar characters is effectively avoided. The conventional CRNN network compresses the height of the input image to 1 to obtain a feature map of 1 × W × C, but the feature map with the height of 1 has insufficient information retained in the dimension, and characters such as Y and V close to the dimension are easily identified. According to the scheme, the Resnet101 compresses the 32W input image into a 6X Y feature map, and richer feature information is reserved on the height, so that the resolution of the feature map is higher in the dimension, and better recognition accuracy is obtained.
In one embodiment, the recognition model includes a page number recognition model and a ticket identification model, the page number identification area is input into the page number recognition model to obtain a recognition result of the page number identification of the questionnaire image to be recognized, and the ticket identification area is input into the ticket identification model to obtain a recognition result of the ticket identification of the questionnaire image to be recognized. Considering that when a page number is identified in an actual scene, the page number represents the number of pages (not the number of printed copies) contained in a form of a style, and the value of the page number is between 1 and 99, so that the identification problem of one page number can be converted into a 99 classification problem of 1 to 99, for example: the number 10 is identified as 1 and 0 in the conventional OCR scheme, and the embodiment considers 10 as a whole and directly corresponds to the 10 th class in the classes 1-99, which has the advantage of converting the OCR problem into a classification problem, so that the CTC Loss in CRNN can be eliminated, and the Cross entry Loss is adopted to avoid the common multi-word or missing-word problem of the CTC Loss (for example, 10 is identified as 110 or 0).
Characters such as the number 1 and the letter I, the number 0 and the letter O in the page number and the ticket are easy to be confused, and in order to avoid such problems, the present embodiment adopts a ticket identification model, i.e., a 26 classification model of a to Z, so as to avoid the problem of easy confusion. Because the data of the page number and the ticket type do not need context information, the page number identification model and the ticket type identification model in the embodiment eliminate a BilSTM module in the CRNN in consideration of the model parameter quantity and the identification speed, so that the model identification speed is superior to that of the traditional CRNN.
The convolutional recurrent neural network model was trained using an Adam optimizer with a learning rate of 0.001. And (3) adopting a cross entropy loss function, setting the batch size to be 64, wherein the batch size is the number of samples trained in the network sent each time, generating a loss value by forward propagation during training, updating the network weight in the direction of reducing the loss value by a back propagation algorithm, and continuously training until the network converges.
And positioning the evaluation item and the evaluation individual corresponding to the input form image according to the identification result of the ticket type and the page number. The method comprises the steps of aligning a questionnaire image to be identified with a questionnaire template image, segmenting a content filling area according to an evaluation item of the questionnaire template image and coordinate information of an evaluation individual to obtain a plurality of evaluation areas, wherein a fillable area with the minimum pixel value in the evaluation area is a filled area due to the fact that grading is carried out in a blacking or other coating mode, obtaining coordinates corresponding to the filled area of the image to be identified according to the coordinate position of the filled area of the questionnaire template image, and accordingly obtaining a result of the evaluation area and feeding the result back to a user. For example: ten areas to be filled with the evaluation areas [1] to [10], the fillable areas [1] to [10] in the questionnaire image to be identified are extracted according to the position information of the questionnaire template image, then the pixel values in the ten fillable areas [1] to [10] are calculated, and the fillable area with the minimum sum of the pixel values is the filled area because the fillable area is blackened.
Further, whether the plurality of evaluation areas have filling missing areas or not can be judged by utilizing a first preset judgment rule, whether the plurality of evaluation areas have filling multi-areas or not can be judged by utilizing a second preset judgment rule, and if the filling missing areas or the filling multi-areas exist, the filling missing areas or the filling multi-areas are added to the result of the evaluation area. For the case of missing fill, a value of (sum of pixels in pure white region-sum of pixels in filled region minimum sum of pixels)/sum of pixels in pure white region is calculated, and when the value is smaller than a third preset threshold, it is determined that the region belongs to an abnormal missing fill, and the third preset threshold can be set to 0.1.
Two values are calculated for the cases of abnormal contamination or multiple fills, respectively: (pure white area pixel sum-fill area minimum pixel sum)/pure white area pixel sum, (pure white area pixel sum-fill area sub-minimum pixel sum)/pure white area pixel sum. If the two values are both larger than a fourth preset threshold value (0.1), judging that the area is polluted or is in an abnormal condition of multiple filling. After the abnormal filling area is found, the system can prompt a user to indicate the image number with the abnormal filling, the user can manually intervene to eliminate the abnormal value, and the identification accuracy is guaranteed.
After the abnormal condition is eliminated, the grading identification results of all the filling areas are respectively corresponding to corresponding individuals and items according to the identification results of the ticket types and the page numbers, and the classified results which are not counted are output to a table together with the abnormal condition detection prompt, so that the user can preliminarily browse the results and backup and retain the results.
Further, the questionnaire image recognition program further includes an analysis module: and performing statistical analysis on the original recognition result to generate a horizontal statistical result and a longitudinal statistical result, or generating a visualization chart such as a bar chart line chart and the like and then feeding back the visualization chart to the user.
After the original recognition result is obtained, the statistical result can be automatically calculated, including statistical sum, average score, weighted average score, highest score, lowest score and the like, and visual charts such as a bar chart, a line chart and the like are generated, so that a user can conveniently browse the result. Compared with the method that the user manually processes the original recognition result, the time for designing the structure of the statistical table head and the statistical table and performing operations such as addition, averaging and the like is saved. If the user carries out the same questionnaire evaluation activity for many times, the statistical result comparison of the activities in multiple periods can be automatically generated by the content stored in the current period database, wherein the statistical result comparison comprises the variation trend of the statistical sum, the average score, the weighted average score, the highest score and the lowest score along with the time, and the statistical result comparison from all aspects of the horizontal direction and the vertical direction is convenient for the user.
In addition, the invention also provides a questionnaire image recognition method. Fig. 3 is a schematic flow chart of a method of an embodiment of the questionnaire image recognition method of the present invention. When processor 12 of electronic device 1 executes questionnaire image recognition program 10 stored in memory 11, the following steps of the questionnaire image recognition method are implemented:
step S10: receiving a request for identifying the questionnaire image sent by a user, and acquiring the questionnaire image to be identified in the request.
In this embodiment, the questionnaire image to be identified, which is captured by the user through the scanner or the mobile phone, can be received, and the system provided by the invention has strong robustness and low requirements on illumination and capturing angles, so that the mode and environment for the user to capture the questionnaire image to be identified are not limited.
The questionnaire image to be identified can comprise a primary table or a secondary table, and the structure of the primary table is as follows: the single-class individual multi-evaluation item can be scored in a single table, and the structure of a secondary table is as follows: multiple evaluation items for multiple classes of individuals can be scored in a single table. In the questionnaire tabulation process of the embodiment, a user can input items to be evaluated and individuals to be evaluated according to requirements in a questionnaire style, and ticket identification, for example, named by English capital letters A-Z, can be added in the upper right corner of a table in order to meet the statistical requirements of the same user on different groups to be evaluated in a classified manner. The table can be stored as a picture or printed for use after being automatically generated, and a JSON format input evaluation information file can be stored for subsequent identification statistics and next reuse. The storage mode can also facilitate the situation that tabulation and recognition statistics work are processed by different people respectively, and a user can carry out subsequent function operation only by opening the stored file in the JSON format by using the system.
Step S20: and performing binarization processing on the questionnaire image to be identified and a preset questionnaire template image, and performing processing in a preset mode on the questionnaire image to be identified and the questionnaire template image after binarization processing.
In this embodiment, the size of the questionnaire image to be recognized is adjusted to an image with the same size as the preset questionnaire template image, adaptive binarization processing is performed on the image to be recognized and the questionnaire template image by using a cv2.thresh _ OTSU method, and the image is converted into pixel values of 0 and 255, so that interferences of different illumination conditions are removed, and then preset mode processing is performed on the questionnaire image to be recognized and the questionnaire template image after binarization processing, wherein the preset mode processing includes performing corrosion processing on a check image adopting (5, 5), and interferences such as noise points and the like can be removed. The preprocessing operation enables the subsequent image matching to be more robust and not easily affected by illumination and printing ink depth.
Step S30: matching and correcting the processed questionnaire image to be recognized with the questionnaire template image to obtain the questionnaire image to be recognized with a corresponding coordinate position of the questionnaire template image, obtaining the coordinates of the ticket identification area, the page identification area and the content filling area of the image to be recognized based on the coordinates of the questionnaire template image, and cutting the ticket identification area, the page identification area and the content filling area of the questionnaire image to be recognized.
In the embodiment, based on the consideration of both stability and speed, the image matching is performed by using the SIFT feature matching method. Specifically, feature points of a questionnaire image to be recognized and a questionnaire template image are analyzed according to an SIFT algorithm, Euclidean distances between each feature point in the questionnaire template image and each feature point in the questionnaire image to be recognized are respectively calculated, two feature points with the minimum distance and the minimum distance in the questionnaire template image and the questionnaire image to be recognized are found out, if the feature point with the minimum distance and the feature point with the minimum distance meet preset conditions, the feature point with the minimum distance and the corresponding feature point in the questionnaire template image are used as a pair of matching feature points, and the questionnaire template image with the logarithm of the matching feature points larger than a first preset value is used as the image matched with the questionnaire image to be recognized. Each feature point in the questionnaire template image and two feature points with the minimum distance and the second minimum distance in the questionnaire image to be identified can be searched by using a knmatch method in cv2.BFMatcher, and the preset conditions are as follows: a value obtained by dividing the distance value of the feature point having the smallest distance by the distance value of the feature point having the next smallest distance is smaller than a second preset value, and the second preset value may be 0.65. If the logarithm of the matching feature points is greater than the first preset value, it indicates that the two image contents include the same feature object, and the first preset value of this embodiment may be set to 4.
After the two images are determined to be matched, the questionnaire image to be recognized and the questionnaire template image are corrected, the coordinates of each feature point of the questionnaire image to be recognized and the questionnaire template image which are successfully matched are respectively obtained, a homography matrix of the template image converted from the image to be matched is obtained through calculation based on the coordinates of each matched feature point, perspective transformation is carried out on the image to be matched based on the homography matrix, and the questionnaire image to be recognized with the questionnaire template image in the corresponding coordinate position is obtained. The method of cv2.RANSAC can be adopted when calculating the homography matrix. The method comprises the steps of obtaining position coordinates of a ticket type identification area, a page number identification area and a content filling area of a questionnaire image to be identified based on coordinates of a ticket type, a page number and a content area of the questionnaire template image, and cutting the ticket type identification area, the page number identification area and the content filling area of the questionnaire image to be identified.
Step S40: inputting the ticket identification area and the page identification area into a pre-trained identification model to obtain identification results of the ticket identification and the page identification of the questionnaire image to be identified, dividing the content filling area based on the identification results and the coordinates of the questionnaire template image to obtain a plurality of evaluation areas, calculating the pixel value of the fillable area of each evaluation area, and taking the fillable area with the minimum pixel value as the result of the evaluation area and feeding back the result to the user.
In this embodiment, the ticket identification area and the page identification area of the questionnaire image to be recognized are input into a recognition model trained in advance based on a convolutional recurrent neural network, so as to obtain the recognition results of the ticket identification and the page identification of the questionnaire image to be recognized. The table type is divided into 26 letters from A to Z, and the table page numbers are numbers of 1, 2, 3 and the like. The CNN module based on VGG16 in the CRNN is replaced by Resnet101, and the Shotcut structure in Resnet can effectively prevent gradient disappearance, so that character features with higher distinguishing degree can be extracted by adopting a deeper network structure, and the problem of identification errors (such as I and L) of similar characters is effectively avoided. The conventional CRNN network compresses the height of the input image to 1 to obtain a feature map of 1 × W × C, but the feature map with the height of 1 has insufficient information retained in the dimension, and characters such as Y and V close to the dimension are easily identified. According to the scheme, the Resnet101 compresses the 32W input image into a 6X Y feature map, and richer feature information is reserved on the height, so that the resolution of the feature map is higher in the dimension, and better recognition accuracy is obtained.
In one embodiment, the recognition model includes a page number recognition model and a ticket identification model, the page number identification area is input into the page number recognition model to obtain a recognition result of the page number identification of the questionnaire image to be recognized, and the ticket identification area is input into the ticket identification model to obtain a recognition result of the ticket identification of the questionnaire image to be recognized. Considering that when a page number is identified in an actual scene, the page number represents the number of pages (not the number of printed copies) contained in a form of a style, and the value of the page number is between 1 and 99, so that the identification problem of one page number can be converted into a 99 classification problem of 1 to 99, for example: the number 10 is identified as 1 and 0 in the conventional OCR scheme, and the embodiment considers 10 as a whole and directly corresponds to the 10 th class in the classes 1-99, which has the advantage of converting the OCR problem into a classification problem, so that the CTC Loss in CRNN can be eliminated, and the Cross entry Loss is adopted to avoid the common multi-word or missing-word problem of the CTC Loss (for example, 10 is identified as 110 or 0).
Characters such as the number 1 and the letter I, the number 0 and the letter O in the page number and the ticket are easy to be confused, and in order to avoid such problems, the present embodiment adopts a ticket identification model, i.e., a 26 classification model of a to Z, so as to avoid the problem of easy confusion. Because the data of the page number and the ticket type do not need context information, the page number identification model and the ticket type identification model in the embodiment eliminate a BilSTM module in the CRNN in consideration of the model parameter quantity and the identification speed, so that the model identification speed is superior to that of the traditional CRNN.
The convolutional recurrent neural network model was trained using an Adam optimizer with a learning rate of 0.001. And (3) adopting a cross entropy loss function, setting the batch size to be 64, wherein the batch size is the number of samples trained in the network sent each time, generating a loss value by forward propagation during training, updating the network weight in the direction of reducing the loss value by a back propagation algorithm, and continuously training until the network converges.
And positioning the evaluation item and the evaluation individual corresponding to the input form image according to the identification result of the ticket type and the page number. The method comprises the steps of aligning a questionnaire image to be identified with a questionnaire template image, segmenting a content filling area according to an evaluation item of the questionnaire template image and coordinate information of an evaluation individual to obtain a plurality of evaluation areas, wherein a fillable area with the minimum pixel value in the evaluation area is a filled area due to the fact that grading is carried out in a blacking or other coating mode, obtaining coordinates corresponding to the filled area of the image to be identified according to the coordinate position of the filled area of the questionnaire template image, and accordingly obtaining a result of the evaluation area and feeding the result back to a user. For example: ten areas to be filled with the evaluation areas [1] to [10], the fillable areas [1] to [10] in the questionnaire image to be identified are extracted according to the position information of the questionnaire template image, then the pixel values in the ten fillable areas [1] to [10] are calculated, and the fillable area with the minimum sum of the pixel values is the filled area because the fillable area is blackened.
Further, whether the plurality of evaluation areas have filling missing areas or not can be judged by utilizing a first preset judgment rule, whether the plurality of evaluation areas have filling multi-areas or not can be judged by utilizing a second preset judgment rule, and if the filling missing areas or the filling multi-areas exist, the filling missing areas or the filling multi-areas are added to the result of the evaluation area. For the case of missing fill, a value of (sum of pixels in pure white region-sum of pixels in filled region minimum sum of pixels)/sum of pixels in pure white region is calculated, and when the value is smaller than a third preset threshold, it is determined that the region belongs to an abnormal missing fill, and the third preset threshold can be set to 0.1.
Two values are calculated for the cases of abnormal contamination or multiple fills, respectively: (pure white area pixel sum-fill area minimum pixel sum)/pure white area pixel sum, (pure white area pixel sum-fill area sub-minimum pixel sum)/pure white area pixel sum. If the two values are both larger than a fourth preset threshold value (0.1), judging that the area is polluted or is in an abnormal condition of multiple filling. After the abnormal filling area is found, the system can prompt a user to indicate the image number with the abnormal filling, the user can manually intervene to eliminate the abnormal value, and the identification accuracy is guaranteed.
After the abnormal condition is eliminated, the grading identification results of all the filling areas are respectively corresponding to corresponding individuals and items according to the identification results of the ticket types and the page numbers, and the classified results which are not counted are output to a table together with the abnormal condition detection prompt, so that the user can preliminarily browse the results and backup and retain the results.
Further, the questionnaire image recognition method further comprises the analysis step of: and performing statistical analysis on the original recognition result to generate a horizontal statistical result and a longitudinal statistical result, or generating a visualization chart such as a bar chart line chart and the like and then feeding back the visualization chart to the user.
After the original recognition result is obtained, the statistical result can be automatically calculated, including statistical sum, average score, weighted average score, highest score, lowest score and the like, and visual charts such as a bar chart, a line chart and the like are generated, so that a user can conveniently browse the result. Compared with the method that the user manually processes the original recognition result, the time for designing the structure of the statistical table head and the statistical table and performing operations such as addition, averaging and the like is saved. If the user carries out the same questionnaire evaluation activity for many times, the statistical result comparison of the activities in multiple periods can be automatically generated by the content stored in the current period database, wherein the statistical result comparison comprises the variation trend of the statistical sum, the average score, the weighted average score, the highest score and the lowest score along with the time, and the statistical result comparison from all aspects of the horizontal direction and the vertical direction is convenient for the user.
Furthermore, the embodiment of the present invention also provides a computer-readable storage medium, which may be any one or any combination of a hard disk, a multimedia card, an SD card, a flash memory card, an SMC, a Read Only Memory (ROM), an Erasable Programmable Read Only Memory (EPROM), a portable compact disc read only memory (CD-ROM), a USB memory, and the like. Included in the computer-readable storage medium is a questionnaire image recognition program 10, which questionnaire image recognition program 10 when executed by a processor performs the following:
a receiving step: receiving a request for identifying a questionnaire image sent by a user, and acquiring the questionnaire image to be identified in the request;
a pretreatment step: performing binarization processing on the questionnaire image to be identified and a preset questionnaire template image, and performing processing in a preset mode on the questionnaire image to be identified and the questionnaire template image after binarization processing;
matching: matching and correcting the processed questionnaire image to be recognized with the questionnaire template image to obtain a questionnaire image to be recognized with a corresponding coordinate position of the questionnaire template image, obtaining coordinates of a ticket identification area, a page identification area and a content filling area of the image to be recognized based on the coordinates of the questionnaire template image, and cutting the ticket identification area, the page identification area and the content filling area of the questionnaire image to be recognized; and
an identification step: inputting the ticket identification area and the page identification area into a pre-trained identification model to obtain identification results of the ticket identification and the page identification of the questionnaire image to be identified, dividing the content filling area based on the identification results and the coordinates of the questionnaire template image to obtain a plurality of evaluation areas, calculating the pixel value of the fillable area of each evaluation area, and taking the fillable area with the minimum pixel value as the result of the evaluation area and feeding back the result to the user.
The specific implementation of the computer readable storage medium of the present invention is substantially the same as the specific implementation of the above questionnaire image recognition method, and will not be described herein again.
It should be noted that the above-mentioned numbers of the embodiments of the present invention are merely for description, and do not represent the merits of the embodiments. And the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, apparatus, article, or method. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, apparatus, article, or method that includes the element.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solution of the present invention essentially or contributing to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) as described above and includes several instructions for enabling a terminal device (such as a mobile phone, a computer, an electronic device, or a network device) to execute the method according to the embodiments of the present invention.
The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (10)
1. A questionnaire image recognition method is applied to an electronic device, and is characterized by comprising the following steps:
a receiving step: receiving a request for identifying a questionnaire image sent by a user, and acquiring the questionnaire image to be identified in the request;
a pretreatment step: performing binarization processing on the questionnaire image to be identified and a preset questionnaire template image, and performing processing in a preset mode on the questionnaire image to be identified and the questionnaire template image after binarization processing;
matching: matching and correcting the processed questionnaire image to be recognized with the questionnaire template image to obtain a questionnaire image to be recognized with a corresponding coordinate position of the questionnaire template image, obtaining coordinates of a ticket identification area, a page identification area and a content filling area of the image to be recognized based on the coordinates of the questionnaire template image, and cutting the ticket identification area, the page identification area and the content filling area of the questionnaire image to be recognized; and
an identification step: inputting the ticket identification area and the page identification area into a pre-trained identification model to obtain identification results of the ticket identification and the page identification of the questionnaire image to be identified, dividing the content filling area based on the identification results and the coordinates of the questionnaire template image to obtain a plurality of evaluation areas, calculating the pixel value of the fillable area of each evaluation area, and taking the fillable area with the minimum pixel value as the result of the evaluation area and feeding back the result to the user.
2. The questionnaire image recognition method of claim 1, wherein the matching step comprises:
analyzing the feature points of the questionnaire image to be identified and the questionnaire template image according to an SIFT algorithm, and respectively calculating the Euclidean distance between each feature point in the questionnaire template image and each feature point in the questionnaire image to be identified;
searching out two feature points with the minimum distance and the minimum distance in the questionnaire template image from each feature point in the questionnaire template image, and if the feature point with the minimum distance and the feature point with the minimum distance meet preset conditions, taking the feature point with the minimum distance and the corresponding feature point in the questionnaire template image as a pair of matched feature points;
and taking the questionnaire template image with the logarithm of the matching feature points larger than a first preset value as the image matched with the questionnaire image to be identified.
3. The questionnaire image recognition method of claim 2, wherein the matching step comprises:
respectively obtaining coordinates of matching feature points of the matched questionnaire image to be identified and the questionnaire template image, calculating to obtain a homography matrix for converting the image to be matched into the template image based on the coordinates of the matching feature points, and carrying out perspective transformation on the image to be matched based on the homography matrix to obtain the questionnaire image to be identified with the questionnaire template image having a corresponding coordinate position.
4. The questionnaire image recognition method of claim 1, wherein the recognition step comprises:
the recognition model comprises a page number recognition model and a ticket identification model, the page number identification area is input into the page number recognition model to obtain a recognition result of the page number identification of the questionnaire image to be recognized, and the ticket identification area is input into the ticket identification model to obtain a recognition result of the ticket identification of the questionnaire image to be recognized.
5. The questionnaire image recognition method of claim 1, wherein the recognizing step further comprises:
judging whether the plurality of evaluation areas have filling missing areas or not by using a first preset judgment rule, judging whether the plurality of evaluation areas have filling areas or not by using a second preset judgment rule, and if the filling missing areas or the filling areas exist, adding the filling missing areas or the filling areas to the result of the evaluation area.
6. An electronic device, comprising a memory and a processor, wherein a questionnaire image recognition program is stored in the memory, and the questionnaire image recognition program is executed by the processor, and the following steps are implemented:
a receiving step: receiving a request for identifying a questionnaire image sent by a user, and acquiring the questionnaire image to be identified in the request;
a pretreatment step: performing binarization processing on the questionnaire image to be identified and a preset questionnaire template image, and performing processing in a preset mode on the questionnaire image to be identified and the questionnaire template image after binarization processing;
matching: matching and correcting the processed questionnaire image to be recognized with the questionnaire template image to obtain a questionnaire image to be recognized with a corresponding coordinate position of the questionnaire template image, obtaining coordinates of a ticket identification area, a page identification area and a content filling area of the image to be recognized based on the coordinates of the questionnaire template image, and cutting the ticket identification area, the page identification area and the content filling area of the questionnaire image to be recognized; and
an identification step: inputting the ticket identification area and the page identification area into a pre-trained identification model to obtain identification results of the ticket identification and the page identification of the questionnaire image to be identified, dividing the content filling area based on the identification results and the coordinates of the questionnaire template image to obtain a plurality of evaluation areas, calculating the pixel value of the fillable area of each evaluation area, and taking the fillable area with the minimum pixel value as the result of the evaluation area and feeding back the result to the user.
7. The electronic device of claim 6, wherein the matching step comprises:
analyzing the feature points of the questionnaire image to be identified and the questionnaire template image according to an SIFT algorithm, and respectively calculating the Euclidean distance between each feature point in the questionnaire template image and each feature point in the questionnaire image to be identified;
searching out two feature points with the minimum distance and the minimum distance in the questionnaire template image from each feature point in the questionnaire template image, and if the feature point with the minimum distance and the feature point with the minimum distance meet preset conditions, taking the feature point with the minimum distance and the corresponding feature point in the questionnaire template image as a pair of matched feature points;
and taking the questionnaire template image with the logarithm of the matching feature points larger than a first preset value as the image matched with the questionnaire image to be identified.
8. The electronic device of claim 7, wherein the matching step comprises:
respectively obtaining coordinates of matching feature points of the matched questionnaire image to be identified and the questionnaire template image, calculating to obtain a homography matrix for converting the image to be matched into the template image based on the coordinates of the matching feature points, and carrying out perspective transformation on the image to be matched based on the homography matrix to obtain the questionnaire image to be identified with the questionnaire template image having a corresponding coordinate position.
9. The electronic device of claim 6, wherein the identifying step comprises:
the recognition model comprises a page number recognition model and a ticket identification model, the page number identification area is input into the page number recognition model to obtain a recognition result of the page number identification of the questionnaire image to be recognized, and the ticket identification area is input into the ticket identification model to obtain a recognition result of the ticket identification of the questionnaire image to be recognized.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium includes a questionnaire image recognition program, and the questionnaire image recognition program, when executed by a processor, implements the steps of the questionnaire image recognition method according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010316545.9A CN111553334A (en) | 2020-04-21 | 2020-04-21 | Questionnaire image recognition method, electronic device, and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010316545.9A CN111553334A (en) | 2020-04-21 | 2020-04-21 | Questionnaire image recognition method, electronic device, and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111553334A true CN111553334A (en) | 2020-08-18 |
Family
ID=72007574
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010316545.9A Withdrawn CN111553334A (en) | 2020-04-21 | 2020-04-21 | Questionnaire image recognition method, electronic device, and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111553334A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112101336A (en) * | 2020-09-09 | 2020-12-18 | 杭州测质成科技有限公司 | Intelligent data acquisition mode based on computer vision |
| CN112417177A (en) * | 2020-12-10 | 2021-02-26 | 清研灵智信息咨询(北京)有限公司 | Knowledge graph construction method based on investigation data |
| CN112818765A (en) * | 2021-01-18 | 2021-05-18 | 中科院成都信息技术股份有限公司 | Image filling identification method, device, system and storage medium |
| CN112949455A (en) * | 2021-02-26 | 2021-06-11 | 武汉天喻信息产业股份有限公司 | Value-added tax invoice identification system and method |
| CN113377356A (en) * | 2021-06-11 | 2021-09-10 | 四川大学 | Method, device, equipment and medium for generating user interface prototype code |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201741154U (en) * | 2010-02-08 | 2011-02-09 | 巩青歌 | System for recognizing and summarizing assessment questionnaire |
| CN108446621A (en) * | 2018-03-14 | 2018-08-24 | 平安科技（深圳）有限公司 | Bank slip recognition method, server and computer readable storage medium |
| CN109558844A (en) * | 2018-11-30 | 2019-04-02 | 厦门商集网络科技有限责任公司 | The method and apparatus of self-defined template discrimination is promoted based on image normalization |
| CN109635796A (en) * | 2018-11-20 | 2019-04-16 | 泰康保险集团股份有限公司 | Recognition methods, device and the equipment of questionnaire |
| CN110263694A (en) * | 2019-06-13 | 2019-09-20 | 泰康保险集团股份有限公司 | A kind of bank slip recognition method and device |
| CN110532938A (en) * | 2019-08-27 | 2019-12-03 | 海南阿凡题科技有限公司 | Papery operation page number recognition methods based on Faster-RCNN |
-
2020
- 2020-04-21 CN CN202010316545.9A patent/CN111553334A/en not_active Withdrawn
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201741154U (en) * | 2010-02-08 | 2011-02-09 | 巩青歌 | System for recognizing and summarizing assessment questionnaire |
| CN108446621A (en) * | 2018-03-14 | 2018-08-24 | 平安科技（深圳）有限公司 | Bank slip recognition method, server and computer readable storage medium |
| CN109635796A (en) * | 2018-11-20 | 2019-04-16 | 泰康保险集团股份有限公司 | Recognition methods, device and the equipment of questionnaire |
| CN109558844A (en) * | 2018-11-30 | 2019-04-02 | 厦门商集网络科技有限责任公司 | The method and apparatus of self-defined template discrimination is promoted based on image normalization |
| CN110263694A (en) * | 2019-06-13 | 2019-09-20 | 泰康保险集团股份有限公司 | A kind of bank slip recognition method and device |
| CN110532938A (en) * | 2019-08-27 | 2019-12-03 | 海南阿凡题科技有限公司 | Papery operation page number recognition methods based on Faster-RCNN |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112101336A (en) * | 2020-09-09 | 2020-12-18 | 杭州测质成科技有限公司 | Intelligent data acquisition mode based on computer vision |
| CN112417177A (en) * | 2020-12-10 | 2021-02-26 | 清研灵智信息咨询(北京)有限公司 | Knowledge graph construction method based on investigation data |
| CN112818765A (en) * | 2021-01-18 | 2021-05-18 | 中科院成都信息技术股份有限公司 | Image filling identification method, device, system and storage medium |
| CN112818765B (en) * | 2021-01-18 | 2023-09-19 | 中科院成都信息技术股份有限公司 | Image filling identification method, device and system and storage medium |
| CN112949455A (en) * | 2021-02-26 | 2021-06-11 | 武汉天喻信息产业股份有限公司 | Value-added tax invoice identification system and method |
| CN112949455B (en) * | 2021-02-26 | 2024-04-05 | 武汉天喻信息产业股份有限公司 | Value-added tax invoice recognition system and method |
| CN113377356A (en) * | 2021-06-11 | 2021-09-10 | 四川大学 | Method, device, equipment and medium for generating user interface prototype code |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111553334A (en) | Questionnaire image recognition method, electronic device, and storage medium | |
| CN111476227B (en) | Target field identification method and device based on OCR and storage medium | |
| CN107403128B (en) | Article identification method and device | |
| CN107689070B (en) | Chart data structured extraction method, electronic device and computer-readable storage medium | |
| CN110675940A (en) | Pathological image labeling method and device, computer equipment and storage medium | |
| CN111695439A (en) | Image structured data extraction method, electronic device and storage medium | |
| CN110942004A (en) | Handwriting recognition method and device based on neural network model and electronic equipment | |
| CN111340023B (en) | Text recognition method and device, electronic equipment and storage medium | |
| CN107977658B (en) | Image character area identification method, television and readable storage medium | |
| CN111291753B (en) | Text recognition method and device based on image and storage medium | |
| CN111325104A (en) | Text recognition method, device and storage medium | |
| CN111310426A (en) | Form format recovery method and device based on OCR and storage medium | |
| CN114092938B (en) | Image recognition processing method and device, electronic equipment and storage medium | |
| US9396389B2 (en) | Techniques for detecting user-entered check marks | |
| CN107403179B (en) | Registration method and device for article packaging information | |
| JP6435934B2 (en) | Document image processing program, image processing apparatus and character recognition apparatus using the program | |
| EP2816504A1 (en) | Character-extraction method and character-recognition device and program using said method | |
| CN114049540A (en) | Method, device, equipment and medium for detecting marked image based on artificial intelligence | |
| CN114005126A (en) | Table reconstruction method and device, computer equipment and readable storage medium | |
| WO2021143058A1 (en) | Image-based information comparison method, apparatus, electronic device, and computer-readable storage medium | |
| CN113569677A (en) | Paper test report generation method based on scanning piece | |
| CN112364857A (en) | Image recognition method and device based on numerical extraction and storage medium | |
| CN107688788B (en) | Document chart extraction method, electronic device and computer readable storage medium | |
| CN115457585A (en) | Processing method and device for homework correction, computer equipment and readable storage medium | |
| US20130330005A1 (en) | Electronic device and character recognition method for recognizing sequential code |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20200818 |
|
| WW01 | Invention patent application withdrawn after publication |