Method, system and device for extracting key information from resume

Abstract

The present invention relates to the technical field of resume image recognition, and is a resume key information extraction method, system and device, comprising: receiving a resume image, performing grayscale processing and correction operations on the resume image to obtain an initial image, obtaining multiple windows to be filled based on a pre-built resume template, and confirming the keyword identifiers of the windows to be filled, constructing resume information data based on the multiple keyword identifiers and the multiple windows to be filled, obtaining a matching company set of the resume image based on the resume information data, extracting matching companies from the matching company set in sequence, and performing the following operations on the extracted matching companies: obtaining a receiving unit of the extracted matching company, sending the resume image to the receiving unit, and completing the extraction of resume key information of the resume image after the receiving unit confirms the receipt of the resume image. The present invention can realize the extraction of key information from a resume image and send the resume to a matching company that meets the conditions according to the key information.

Description

Technical Field

[0001] The present invention relates to the technical field of resume image recognition, and in particular to a resume key information extraction method, system and device. Background Art

[0002] During the recruitment process, recruiting companies need to manually screen a large number of resumes. Usually, resumes are mostly sent in the form of images, so a method that can identify key information in resume images is needed.

[0003] At present, the method of identifying resume images is mostly through human eye recognition, that is, after receiving a large number of resumes, the recruitment company manually checks the resumes one by one, selects the majors and colleges that meet the corresponding positions of the recruitment company, and then manually records the information in the resume image.

[0004] Although the above method can realize the screening of a large number of resumes, it is easy to make information recognition errors in the manual recording process and consumes a lot of manpower. Therefore, a more intelligent method that can automatically recognize data in resume images is needed. Summary of the invention

[0005] The present invention provides a resume key information extraction method, system and device, the main purpose of which is to extract key information from a resume image and send the resume to a matching company that meets the requirements according to the key information.

[0006] To achieve the above purpose, the present invention provides a method for extracting key information from a resume, comprising:

[0007] Receive a resume image, perform grayscale processing and correction operations on the resume image to obtain an initial image, obtain a marker point coordinate set of the initial image based on a pre-built resume template and the initial image, wherein the marker point coordinate set includes a plurality of coordinate sets to be filled in, extract the coordinate sets to be filled in sequence from the marker point coordinate set based on the initial image, and perform the following operations on the extracted coordinate sets to be filled in:

[0008] Obtaining a window to be filled corresponding to the extracted coordinate set to be filled, and confirming a keyword identifier of the window to be filled, obtaining a pixel sum of the extracted window to be filled, and after confirming that the window to be filled has been filled based on the pixel sum, dividing the window to be filled using a pre-built text segmentation method to obtain a keyword image sequence;

[0009] The step of obtaining the pixel sum of the extracted window to be filled, and confirming that the window to be filled has been filled based on the pixel sum, comprises:

[0010] Using a pre-built binarization method to perform a binarization operation on the window to be filled, obtaining a binary window image, obtaining multiple pixel values ​​of the binary window image, and calculating a pixel sum according to the multiple pixel values;

[0011] Comparing the pixel sum with a preset sum threshold;

[0012] If the pixel sum is less than or equal to the sum threshold, it is confirmed that the window to be filled is not filled, and after performing an unfilled marking operation on the window to be filled, an unfilled marking window is obtained, and it is confirmed that the window to be filled is filled based on the unfilled marking window;

[0013] If the pixel sum is greater than the sum threshold, it is confirmed that the window to be filled has been filled;

[0014] Using a pre-built text recognition method to recognize the keyword image sequence to obtain key text, and using a keyword identifier to perform an identification operation on the key text to obtain unit key data;

[0015] Summarize key data of the unit to obtain resume information data corresponding to the resume image, and obtain a matching company set of the resume image based on the resume information data;

[0016] Wherein, the step of obtaining a matching company set of the resume image based on the resume information data includes:

[0017] Constructing a resume matching unit;

[0018] Extracting school data and major data based on the resume information data, obtaining school code and major code based on the school data and major data, calculating a hash value to be matched according to the school code and major code, inputting the hash value to be matched into a resume matching unit, and obtaining a matching company set;

[0019] The construction of the resume matching unit includes:

[0020] Get the matching company set, extract matching companies from the matching company set in sequence, and perform the following operations on the extracted matching companies:

[0021] The target professional set of the extracted matching companies is obtained, target professional sets are extracted in sequence, and the following operations are performed on the extracted target professional sets:

[0022] Obtain a target school set corresponding to the extracted target major, extract target schools from the target school set in sequence, and perform the following operations on the extracted target schools:

[0023] Constructing unit template division rules based on the extracted target colleges, the extracted target majors and the extracted matching companies, and constructing a recruitment identification code according to the unit template division rules;

[0024] Aggregate the recruitment identification codes to obtain an information classification library, and use the information classification library to construct a resume matching unit;

[0025] After confirming that the matching company set is not an empty set, matching companies are extracted from the matching company set in sequence, and the following operations are performed on the extracted matching companies:

[0026] The receiving unit of the extracted matching company is obtained, and the resume image is sent to the receiving unit. After the receiving unit confirms receipt of the resume image, the resume key information of the resume image is extracted.

[0027] Optionally, performing grayscale processing and correction operations on the resume image to obtain an initial image includes:

[0028] Acquire a plurality of pixels based on the resume image, extract pixels from the plurality of pixels in sequence, and perform the following operations on the extracted pixels:

[0029] Obtaining the red channel value, green channel value, and blue channel value of the extracted pixel, inputting the red channel value, green channel value, and blue channel value into a pre-built grayscale conversion formula to obtain a grayscale value, and assigning the grayscale value to the extracted pixel to obtain a grayscale pixel;

[0030] Aggregate the grayscale pixels to get a grayscale image;

[0031] A correction operation is performed on the grayscale image to obtain an initial image.

[0032] Optionally, performing a correction operation on the grayscale image to obtain an initial image includes:

[0033] A grayscale image coordinate system corresponding to the grayscale image is obtained, an image vertical axis of the grayscale image is obtained based on the grayscale image coordinate system, a vertical projection operation is performed on the grayscale image based on the image vertical axis to obtain a plurality of row boundary sets, row boundary sets are sequentially extracted from the plurality of row boundary sets, and the following operations are performed on the extracted row boundary sets:

[0034] Fitting an upper boundary slope based on the extracted row boundary set, and comparing the upper boundary slope with a preset slope threshold;

[0035] If the upper boundary slope is greater than the slope threshold, obtaining a line text image corresponding to the upper boundary slope, and performing a correction operation on the line text image to obtain a corrected line image;

[0036] Otherwise, obtaining a line text image corresponding to the upper boundary slope, and confirming that the line text image is a line image that does not need correction;

[0037] Summarize the corrected row images to obtain a corrected row image set, and summarize the row images that do not need to be corrected to obtain an original row image set;

[0038] After constructing a first correction image based on the corrected row image set and the original row image set, obtaining a correction number according to the first correction image;

[0039] Comparing the correction number with a preset correction number threshold, if the correction number is greater than the correction number threshold, confirming that the first corrected image corresponding to the correction number is the initial image;

[0040] Otherwise, confirm that the first corrected image is a grayscale image, and return to the step of obtaining the grayscale image coordinates corresponding to the grayscale image, until the corrected row image set is an empty set or the correction number is greater than the correction number threshold, to obtain the initial image.

[0041] Optionally, fitting the upper boundary slope based on the extracted row boundary set includes:

[0042] Acquire multiple row upper boundary pixels in the row boundary set, extract the pixel horizontal coordinate of each row upper boundary pixel in turn from the multiple row upper boundary pixels to obtain a pixel horizontal coordinate set, perform a linear fitting operation on the pixel horizontal coordinate set to obtain an upper boundary straight line, and calculate the upper boundary slope based on the upper boundary straight line.

[0043] Optionally, the step of dividing the window to be filled by using a pre-built text segmentation method to obtain a keyword image sequence includes:

[0044] Acquire a binary window image of the window to be filled, acquire a binary coordinate system based on the binary window image, acquire a binary vertical axis based on the binary coordinate system, perform a vertical projection operation on the binary window image according to the binary vertical axis, and obtain a binary row pixel set;

[0045] Dividing the binary row pixel set to obtain a plurality of unit row pixel sets;

[0046] Extracting unit row pixel sets from the plurality of unit row pixel sets in sequence, and performing the following operations on the extracted unit row pixel sets:

[0047] Obtain a binary row image based on the extracted unit row pixel set, obtain a binary horizontal axis based on a binary coordinate system, perform a vertical projection operation on the binary row image based on the binary horizontal axis to obtain a binary vertical pixel set, and divide the binary vertical pixel set using a pre-constructed character division method to obtain a keyword image;

[0048] Summarize keyword images to obtain keyword image sequences.

[0049] Optionally, the using a pre-built binarization method to perform a binarization operation on the window to be filled to obtain a binary window image includes:

[0050] A grayscale set is obtained based on the window to be filled, grayscale values ​​are extracted from the grayscale set in sequence, and the following operations are performed on the extracted grayscale values:

[0051] The extracted grayscale value is removed from the grayscale set to obtain a grayscale value set to be classified. After confirming that one or more pixels corresponding to the extracted grayscale value are pixels of the same type, the grayscale values ​​to be classified are sequentially extracted from the grayscale value set to be classified, and the following operations are performed on the extracted grayscale values ​​to be classified:

[0052] Compare the extracted grayscale value to be classified with the extracted grayscale level value;

[0053] If the extracted grayscale value to be classified is less than or equal to the extracted grayscale level value, one or more pixels corresponding to the extracted grayscale value to be classified are confirmed as first-class pixels; otherwise, one or more pixels corresponding to the extracted grayscale value to be classified are confirmed as second-class pixels;

[0054] The first-class pixels and the second-class pixels are respectively summarized to obtain the first-class pixel set and the second-class pixel set, and the first-class pixel probability and the second-class pixel probability are respectively calculated based on the first-class pixel set and the second-class pixel set, wherein the first-class pixel probability is:

[0055]

[0056] in, represents the probability of a class of pixels, Indicates the number of grayscale values ​​to be classified. Indicates the number of pixels corresponding to the grayscale value to be divided, Indicates the total number of pixels in the window to be filled;

[0057] The maximum inter-class variance of the extracted grayscale values ​​is calculated based on the probability of the first class pixels and the probability of the second class pixels, wherein the calculation formula of the maximum inter-class variance is:

[0058]

[0059] in, represents the maximum between-class variance, represents the probability of a class of pixels, represents the probability of two types of pixels, Represents the mean gray value of pixels in a class of pixels. Represents the mean grayscale value of all pixels corresponding to the window to be filled, Represents the mean of the grayscale values ​​of the pixels in the second type of pixel set;

[0060] The maximum inter-class variances are summarized to obtain a maximum inter-class variance set, an optimal partitioning threshold is determined based on the maximum inter-class variance set, and a binarization operation is performed on the window to be filled according to the optimal partitioning threshold to obtain a binary window image.

[0061] To achieve the above object, the present invention also provides a resume key information extraction system, comprising:

[0062] An image preprocessing module is used to receive a resume image, perform grayscale processing and correction operations on the resume image, and obtain an initial image;

[0063] The unit key data acquisition module is used to acquire the identification point coordinate set of the initial image based on the pre-built resume template and the initial image, wherein the identification point coordinate set includes a plurality of coordinate sets to be filled in, and the coordinate sets to be filled in are sequentially extracted from the identification point coordinate set based on the initial image, and the following operations are performed on the extracted coordinate sets to be filled in: obtaining the window to be filled corresponding to the extracted coordinate set to be filled in, and confirming the keyword identifier of the window to be filled in, obtaining the pixel sum of the extracted window to be filled in, confirming that the window to be filled in has been filled in based on the pixel sum, dividing the window to be filled in by using the pre-built text segmentation method, obtaining a keyword image sequence, recognizing the keyword image sequence by using the pre-built text recognition method, obtaining the key text, and using the keyword identifier to identify the key text. The key text performs a marking operation to obtain unit key data, wherein the obtaining of the pixel sum of the extracted window to be filled, and after confirming that the window to be filled has been filled based on the pixel sum, comprises: performing a binarization operation on the window to be filled using a pre-built binarization method to obtain a binary window image, obtaining multiple pixel values ​​of the binary window image, calculating the pixel sum according to the multiple pixel values, comparing the pixel sum with a preset sum threshold, if the pixel sum is less than or equal to the sum threshold, confirming that the window to be filled has not been filled, performing a non-filled marking operation on the window to be filled, obtaining an non-filled marking window, confirming that the window to be filled has been filled based on the non-filled marking window, and if the pixel sum is greater than the sum threshold, confirming that the window to be filled has been filled;

[0064] The matching company acquisition module is used to summarize the key data of the unit, obtain the resume information data corresponding to the resume image, and obtain the matching company set of the resume image based on the resume information data, wherein the obtaining the matching company set of the resume image based on the resume information data includes: constructing a resume matching unit, extracting school data and major data based on the resume information data, obtaining the school code and major code based on the school data and major data, calculating the hash value to be matched according to the school code and major code, inputting the hash value to be matched into the resume matching unit, and obtaining the matching company set, wherein the constructing of the resume matching unit includes: obtaining the matching company set, extracting the matching company set from the matching company set according to the school code and major code, and obtaining the matching company set according to the school code and major code. Extract matching companies one by one, and perform the following operations on the extracted matching companies: obtain the target major set of the extracted matching companies, extract target majors from the target major set in sequence, and perform the following operations on the extracted target majors: obtain the target school set corresponding to the extracted target major, extract target schools from the target school set in sequence, and perform the following operations on the extracted target schools: construct unit template division rules based on the extracted target schools, the extracted target majors and the extracted matching companies, construct recruitment identification codes according to the unit template division rules, summarize the recruitment identification codes, obtain an information classification library, and use the information classification library to construct a resume matching unit;

[0065] The information sending module is used to confirm that the matching company set is not an empty set, extract matching companies from the matching company set in sequence, and perform the following operations on the extracted matching companies: obtain the receiving unit of the extracted matching company, send the resume image to the receiving unit, and complete the extraction of resume key information of the resume image after the receiving unit confirms the receipt of the resume image.

[0066] In order to solve the above problem, the present invention further provides an electronic device, the electronic device comprising:

[0067] a memory storing at least one instruction; and

[0068] The processor executes the instructions stored in the memory to implement the resume key information extraction method described above.

[0069] In order to solve the above problem, the present invention also provides a computer-readable storage medium, in which at least one instruction is stored. The at least one instruction is executed by a processor in an electronic device to implement the above-mentioned resume key information extraction method.

[0070] The present invention is to solve the problem described in the background technology. The present invention receives a resume image, performs grayscale processing and correction operations on the resume image, and obtains an initial image. The present invention performs grayscale operations on the resume image to facilitate correction processing. When the characters in the resume image are skewed and distorted, the accuracy of text recognition will be affected. The purpose of correction is to more easily identify the information in the resume image, so the present invention corrects the resume image line by line. Based on a pre-constructed resume template and an initial image, a coordinate set of identification points of an initial image is obtained; based on the initial image, a coordinate set to be filled in is extracted from the coordinate set of identification points in sequence; and the following operations are performed on the extracted coordinate sets to be filled in: a window to be filled corresponding to the extracted coordinate set to be filled is obtained, and a keyword identifier of the window to be filled is confirmed; the pixel sum of the extracted window to be filled is obtained; after confirming that the window to be filled has been filled based on the pixel sum, a pre-constructed text segmentation method is used to divide the window to be filled to obtain a keyword image sequence; a pre-constructed text recognition method is used to recognize the keyword image sequence to obtain key text; a marking operation is performed on the key text using the keyword identifier to obtain unit key data; the present invention detects each window to be filled in. Since the initial image has been obtained, compared with directly extracting text and finding that the window to be filled has not been filled in, and then directly using the pixel sum to judge whether it is filled in, it can judge whether the window to be filled in is filled in more quickly. Summarize the key data of the unit, obtain the resume information data corresponding to the resume image, obtain the matching company set of the resume image based on the resume information data, confirm that the matching company set is not an empty set, extract the matching companies from the matching company set in sequence, and perform the following operations on the extracted matching companies: obtain the receiving unit of the extracted matching company, send the resume image to the receiving unit, and complete the extraction of the resume key information of the resume image after the receiving unit confirms the receipt of the resume image. The present invention extracts the college and major from the resume information, and sends the patent image to the resume matching unit according to the college and major, uses the data in the resume matching unit to match the major and college corresponding to the resume image with the college and major expected by the matching company, and finally sends the resume image to each matching company in the matching company set, that is, the embodiment of the present invention can extract the major and college from the resume image, and send the resume image that meets the recruitment expectations of the matching company to the receiving unit of the matching company, which is convenient for job seekers to send resumes and also conducive to matching companies to receive resumes. Therefore, the present invention extracts key information from the resume image and sends the resume to the matching company that meets the conditions according to the key information. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 A schematic diagram of a flow chart of a resume key information extraction method provided by an embodiment of the present invention;

[0072] Figure 2A functional module diagram of a resume key information extraction system provided by an embodiment of the present invention;

[0073] Figure 3 A schematic diagram of the structure of an electronic device for implementing the resume key information extraction method provided in one embodiment of the present invention.

[0074] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0075] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0076] The embodiment of the present application provides a method for extracting key information from a resume. The execution subject of the method for extracting key information from a resume includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present application. In other words, the method for extracting key information from a resume can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0077] Reference Figure 1 FIG. 1 is a flow chart of a resume key information extraction method provided by an embodiment of the present invention. In this embodiment, the resume key information extraction method includes:

[0078] S1. Receive a resume image, perform grayscale processing and correction operations on the resume image, and obtain an initial image.

[0079] It should be noted that the resume image is an image generated by filling in the resume information of the job seeker, and the initial image is an image after grayscale processing and correction of the resume image.

[0080] Furthermore, performing grayscale processing and correction operations on the resume image to obtain an initial image includes:

[0081] Acquire a plurality of pixels based on the resume image, extract pixels from the plurality of pixels in sequence, and perform the following operations on the extracted pixels:

[0082] Obtaining the red channel value, green channel value, and blue channel value of the extracted pixel, inputting the red channel value, green channel value, and blue channel value into a pre-built grayscale conversion formula to obtain a grayscale value, and assigning the grayscale value to the extracted pixel to obtain a grayscale pixel;

[0083] Aggregate the grayscale pixels to get a grayscale image;

[0084] A correction operation is performed on the grayscale image to obtain an initial image.

[0085] It is understandable that the red channel value, the green channel value and the blue channel value are the values ​​corresponding to the red channel, the green channel and the blue channel in the RGB image, respectively. The grayscale conversion formula is a formula for converting the red channel value, the green channel value and the blue channel value corresponding to the extracted pixel into grayscale values. This technology is prior art and will not be described in detail here. Grayscale pixels are pixels after the extracted pixels are grayscaled.

[0086] Furthermore, performing a correction operation on the grayscale image to obtain an initial image includes:

[0087] A grayscale image coordinate system corresponding to the grayscale image is obtained, an image vertical axis of the grayscale image is obtained based on the grayscale image coordinate system, a vertical projection operation is performed on the grayscale image based on the image vertical axis to obtain a plurality of row boundary sets, row boundary sets are sequentially extracted from the plurality of row boundary sets, and the following operations are performed on the extracted row boundary sets:

[0088] Fitting an upper boundary slope based on the extracted row boundary set, and comparing the upper boundary slope with a preset slope threshold;

[0089] If the upper boundary slope is greater than the slope threshold, obtaining a line text image corresponding to the upper boundary slope, and performing a correction operation on the line text image to obtain a corrected line image;

[0090] Otherwise, obtaining a line text image corresponding to the upper boundary slope, and confirming that the line text image is a line image that does not need correction;

[0091] Summarize the corrected row images to obtain a corrected row image set, and summarize the row images that do not need to be corrected to obtain an original row image set;

[0092] After constructing a first correction image based on the corrected row image set and the original row image set, obtaining a correction number according to the first correction image;

[0093] Comparing the correction number with a preset correction number threshold, if the correction number is greater than the correction number threshold, confirming that the first corrected image corresponding to the correction number is the initial image;

[0094] Otherwise, confirm that the first corrected image is a grayscale image, and return to the step of obtaining the grayscale image coordinates corresponding to the grayscale image, until the corrected row image set is an empty set or the correction number is greater than the correction number threshold, to obtain the initial image.

[0095] It is understandable that the grayscale image coordinate system is the image coordinate system of the grayscale image, and the image coordinate system is a universal coordinate system in image processing, so the horizontal axis, vertical axis and pixel coordinates of the image coordinate system in the embodiment of the present invention are not described in detail. It should be noted that although the names corresponding to different images in the embodiment of the present invention are different, the coordinate systems corresponding to different images are all constructed based on the image coordinate system.

[0096] Furthermore, the vertical axis of the image is the vertical axis in the grayscale image coordinate system, and the vertical projection operation is the operation of projecting the grayscale image along the vertical axis of the image. The row boundary set is the set of pixels in the boundary corresponding to a row of text in the grayscale image. The slope threshold is the maximum slope for which the row of text corresponding to the upper boundary slope is considered not tilted and does not need to be corrected. When the upper boundary slope is greater than the slope threshold, it means that the row of text corresponding to the upper boundary slope is too tilted and needs to be corrected. When the characters in the resume image are skewed and distorted, it will affect the accuracy of text recognition, so the purpose of correction is to make it easier to identify the information in the resume image.

[0097] Specifically, the line text image is an image of a line of text corresponding to the extracted line boundary set in the grayscale image. The corrected line image is the corrected line text image. The line image that does not need to be corrected is an image that does not need to be corrected, that is, the embodiment of the present invention believes that a line of text in the line image that does not need to be corrected is not tilted or distorted. The first correction image is an image spliced ​​by using the corrected line image set and the original line image set, but whether the spliced ​​first correction image has reached a level that does not affect the accuracy of text recognition needs to be further verified. Therefore, in the embodiment of the present invention, a correction number is obtained based on the first correction image. The correction number is the number of times the first correction image is obtained. Each time the grayscale image is corrected to obtain the first correction image, the correction number is increased by one. The correction number threshold is the maximum correction number set in the embodiment of the present invention.

[0098] It is understandable that if the correction number is less than the correction number threshold, the degree to which the first corrected image affects the accuracy of text recognition is re-judged, and the basis for the re-judgment is whether the corrected row image set is an empty set. When the corrected row image set is an empty set, the re-judgment of the first corrected image is stopped, and the current first corrected image is output as the initial image. If the corrected row image set is not an empty set, the first corrected image is constructed according to the corrected row image. If the correction number is greater than the correction number threshold, the correction is stopped.

[0099] It should be noted that the purpose of this step is: if the first corrected image after the correction number threshold still cannot satisfy the condition that the corrected row image set is an empty set, then the loop verification of the first corrected image is terminated to avoid infinite operation of the system. That is, the stopping judgment criterion of this loop step is: if one of the corrected row image set is an empty set or the correction number is greater than the correction number threshold is satisfied, then the loop is stopped.

[0100] Further, fitting the upper boundary slope based on the extracted row boundary set includes:

[0101] Acquire multiple row upper boundary pixels in the row boundary set, extract the pixel horizontal coordinate of each row upper boundary pixel in turn from the multiple row upper boundary pixels to obtain a pixel horizontal coordinate set, perform a linear fitting operation on the pixel horizontal coordinate set to obtain an upper boundary straight line, and calculate the upper boundary slope based on the upper boundary straight line.

[0102] Furthermore, the upper boundary pixel is the pixel where the upper boundary of a line of text in the grayscale image is located. The pixel horizontal coordinate is the horizontal coordinate of the upper boundary pixel in the grayscale image coordinate system. The upper boundary straight line is a straight line fitted by the pixel horizontal coordinate set. Optionally, in an embodiment of the present invention, the upper boundary straight line is fitted using the least squares method and the pixel horizontal coordinate set. The upper boundary slope is the slope of the upper boundary straight line.

[0103] S2. Obtain a set of identification point coordinates of the initial image based on a pre-built resume template and the initial image, wherein the set of identification point coordinates includes a plurality of sets of coordinates to be filled in, and sequentially extract the sets of coordinates to be filled in from the set of identification point coordinates.

[0104] It should be noted that the resume template is a template for filling out a resume. When constructing the resume template, the embodiment of the present invention performs a marking operation on the items to be filled in the resume template. Therefore, the resume template includes multiple coordinate sets to be filled in, and each area corresponding to the coordinate set to be filled in corresponds to an item to be filled in.

[0105] It is understandable that the embodiment of the present invention can migrate the identification point coordinate set corresponding to the resume template to the initial image by comparing the corrected initial image with the resume template. Therefore, the corresponding key information in the initial image can be confirmed through the identification point coordinate set.

[0106] For example, the coordinates corresponding to the name area in the resume template are: (0,0), (20,0), (0,10) and (20,10). Therefore, the area enclosed by (0,0), (20,0), (0,10) and (20,10) in the initial image corresponds to the name area in the resume template. Therefore, the keyword represented by the area enclosed by (0,0), (20,0), (0,10) and (20,10) in the initial image is name.

[0107] S3, and performing the following operations on the extracted coordinate set to be filled: obtaining the window to be filled corresponding to the extracted coordinate set to be filled, and confirming the keyword identifier of the window to be filled, obtaining the pixel sum of the extracted window to be filled, and after confirming that the window to be filled has been filled based on the pixel sum, using a pre-constructed text segmentation method to divide the window to be filled to obtain a keyword image sequence, using a pre-constructed text recognition method to recognize the keyword image sequence to obtain key text, and using the keyword identifier to perform a marking operation on the key text to obtain unit key data;

[0108] Summarize key data of the unit to obtain resume information data corresponding to the resume image, and obtain a matching company set of the resume image based on the resume information data;

[0109] After confirming that the matching company set is not an empty set, matching companies are extracted from the matching company set in sequence, and the following operations are performed on the extracted matching companies:

[0110] The receiving unit of the extracted matching company is obtained, and the resume image is sent to the receiving unit. After the receiving unit confirms receipt of the resume image, the resume key information of the resume image is extracted.

[0111] It is understandable that the keyword identifier is an identifier used to indicate the information corresponding to the window to be filled in. For example, if the window to be filled in needs to be filled in with a name, the keyword identifier identifies the window to be filled in as a name.

[0112] Further, the step of obtaining the pixel sum of the extracted window to be filled, and confirming that the window to be filled has been filled based on the pixel sum, includes:

[0113] Using a pre-built binarization method to perform a binarization operation on the window to be filled, obtaining a binary window image, obtaining multiple pixel values ​​of the binary window image, and calculating a pixel sum according to the multiple pixel values;

[0114] Comparing the pixel sum with a preset sum threshold;

[0115] If the pixel sum is less than or equal to the sum threshold, it is confirmed that the window to be filled is not filled, and after performing an unfilled marking operation on the window to be filled, an unfilled marking window is obtained, and it is confirmed that the window to be filled is filled based on the unfilled marking window;

[0116] If the pixel sum is greater than the sum threshold, it is confirmed that the window to be filled has been filled.

[0117] It is understandable that the binary window image is the image of the extracted window to be filled after binarization. The pixel sum is the cumulative sum of multiple pixel values ​​of the binary window image. The sum threshold is the cumulative threshold of multiple pixel values ​​after the binary window image has been filled. When the pixel sum is less than the sum threshold, it means that the window to be filled corresponding to the binary window image is not filled. Therefore, the key information of the window to be filled cannot be extracted, so it is necessary to mark the window to be filled. In the embodiment of the present invention, the unfilled marking operation is to enter the "unfilled" character in the unfilled window to be filled. The unfilled marking window is a window to be filled with the character "unfilled" in the window to be filled. When the "unfilled" character is entered in the window to be filled that was originally unfilled, an unfilled marking window is obtained, and it is considered that the unfilled marking window has been filled.

[0118] Furthermore, the method of using a pre-built binarization method to perform a binarization operation on the window to be filled to obtain a binary window image includes:

[0119] A grayscale set is obtained based on the window to be filled, grayscale values ​​are extracted from the grayscale set in sequence, and the following operations are performed on the extracted grayscale values:

[0120] The extracted grayscale value is removed from the grayscale set to obtain a grayscale value set to be classified. After confirming that one or more pixels corresponding to the extracted grayscale value are pixels of the same type, the grayscale values ​​to be classified are sequentially extracted from the grayscale value set to be classified, and the following operations are performed on the extracted grayscale values ​​to be classified:

[0121] Compare the extracted grayscale value to be classified with the extracted grayscale level value;

[0122] If the extracted grayscale value to be classified is less than or equal to the extracted grayscale level value, one or more pixels corresponding to the extracted grayscale value to be classified are confirmed as first-class pixels; otherwise, one or more pixels corresponding to the extracted grayscale value to be classified are confirmed as second-class pixels;

[0123] The first-class pixels and the second-class pixels are respectively summarized to obtain the first-class pixel set and the second-class pixel set, and the first-class pixel probability and the second-class pixel probability are respectively calculated based on the first-class pixel set and the second-class pixel set, wherein the first-class pixel probability is:

[0124]

[0125] in, represents the probability of a class of pixels, Indicates the number of grayscale values ​​to be classified. Indicates the number of pixels corresponding to the grayscale value to be divided, Indicates the total number of pixels in the window to be filled;

[0126] The maximum inter-class variance of the extracted grayscale values ​​is calculated based on the probability of the first class pixels and the probability of the second class pixels, wherein the calculation formula of the maximum inter-class variance is:

[0127]

[0128] in, represents the maximum between-class variance, represents the probability of a class of pixels, represents the probability of two types of pixels, Represents the mean gray value of pixels in a class of pixels. Represents the mean grayscale value of all pixels corresponding to the window to be filled, Represents the mean of the grayscale values ​​of the pixels in the second type of pixel set;

[0129] The maximum inter-class variances are summarized to obtain a maximum inter-class variance set, an optimal partitioning threshold is determined based on the maximum inter-class variance set, and a binarization operation is performed on the window to be filled according to the optimal partitioning threshold to obtain a binary window image.

[0130] Furthermore, the gray level set is a set of pixels with the same gray value in the window to be filled, which are classified so that the same gray value is a level. The gray level value is an element in the gray level set, and the gray value set to be divided is a set after the gray level value extracted from the gray level set is removed. Class I pixels are pixels whose corresponding gray value is less than the extracted gray level value. Class II pixels and class II pixel probabilities correspond to Class I pixels and class I pixel probabilities, which will not be repeated here.

[0131] For example, the pixel value in the image corresponding to the window to be filled is: [0, 0, 0, 0, 1, 1, 14, 46, 44], then the gray level set is: [0, 1, 14, 46, 44], the gray level value extracted from the gray level set is 1, so the gray value set to be divided is: [0, 14, 46, 44], and the gray level value corresponding to the first type of pixel set is: [0], and the gray level value corresponding to the second type of pixel set is [0, 14, 46, 44]

[0132] Specifically, the optimal partitioning threshold is the smallest maximum inter-class variance in the maximum inter-class variance set. In the embodiment of the present invention, the window to be filled is partitioned according to the optimal partitioning threshold to obtain a binary window image.

[0133] S4. Divide the window to be filled by using a pre-built text segmentation method to obtain a keyword image sequence.

[0134] Furthermore, the method of dividing the window to be filled by using a pre-built text segmentation method to obtain a keyword image sequence includes:

[0135] Acquire a binary window image of the window to be filled, acquire a binary coordinate system based on the binary window image, acquire a binary vertical axis based on the binary coordinate system, perform a vertical projection operation on the binary window image according to the binary vertical axis, and obtain a binary row pixel set;

[0136] Dividing the binary row pixel set to obtain a plurality of unit row pixel sets;

[0137] Extracting unit row pixel sets from the plurality of unit row pixel sets in sequence, and performing the following operations on the extracted unit row pixel sets:

[0138] Obtain a binary row image based on the extracted unit row pixel set, obtain a binary horizontal axis based on a binary coordinate system, perform a vertical projection operation on the binary row image based on the binary horizontal axis to obtain a binary vertical pixel set, and divide the binary vertical pixel set using a pre-constructed character division method to obtain a keyword image;

[0139] Summarize keyword images to obtain keyword image sequences.

[0140] Further, the binary coordinate system is the coordinate system of the binary window image in the image coordinate system. The binary vertical axis and the binary horizontal axis are the vertical axis and the horizontal axis of the binary coordinate system. The binary row pixel set is the pixel set obtained by projecting the binary window image along the binary vertical axis direction. The vertical projection operation is a prior art.

[0141] Exemplarily, in a binary coordinate system, the set of all pixels corresponding to a ordinate of 0, the set of all pixels corresponding to a ordinate of 1, ..., the set of all pixels corresponding to a ordinate of n is a binary row pixel set. There is a line spacing between a row of text and the next adjacent row of text. Because the part where the line spacing is located is blank, the binary row pixel set can be divided to obtain multiple unit row pixel sets. The unit row pixel set is an upper boundary pixel set and a lower boundary pixel set corresponding to a row of text. Based on the upper boundary pixel set and the lower boundary pixel set, the binary row image corresponding to the unit row pixel set can be confirmed.

[0142] It should be noted that the binary row image is the image corresponding to a line of text in the binary window image, and the binary vertical pixel set is the pixel set obtained by projecting the binary row image along the binary horizontal axis. The keyword image is the image corresponding to a character, and the acquisition method of the keyword image is similar to that of the binary row image, which will not be repeated here.

[0143] Furthermore, the keyword images are arranged in the order of the binary window images to obtain a keyword image sequence.

[0144] S5. Using a pre-built text recognition method to recognize the keyword image sequence to obtain key text, and using a keyword tag to perform a tagging operation on the key text to obtain unit key data.

[0145] It is understandable that the embodiment of the present invention segments the binary window image before acquiring the keyword image sequence, so that each character can be segmented, making it more convenient to recognize the resume image using the text recognition method. The key text is the text contained in the keyword image sequence. The unit key data is the data composed of the key text and the keyword identifier. After the replacement text is identified by the keyword identifier, it is convenient to search.

[0146] S6. Summarize key unit data to obtain resume information data corresponding to the resume image, and obtain a matching company set of the resume image based on the resume information data.

[0147] Further, the obtaining of a matching company set of the resume image based on the resume information data includes:

[0148] Constructing a resume matching unit;

[0149] Based on the resume information data, school data and major data are extracted, based on the school data and major data, school code and major code are obtained, hash value to be matched is calculated according to the school code and major code, and the hash value to be matched is input into the resume matching unit to obtain a matching company set.

[0150] It should be noted that the resume information data is the data of the information included in the resume image. The data corresponding to the exemplary resume information data include: name, gender, date of birth, college, major and contact number. And each unit key data in the resume information data has been identified, so when the company wants to obtain a resume image that meets the recruitment expectations, it can be searched through the resume information data. The resume matching unit is a unit for matching resumes. The resume matching unit can send the resume image to the receiving end of the matching company whose resume meets the needs of the matching company. The college data is the data of the college included in the resume information, the major data is the data of the major included in the resume information data, and the college code and the major code are codes issued by the Ministry of Education of the People's Republic of China. The college code and the major code represent each different college and different major uniquely. The hash value to be matched is a hash value corresponding to the extracted college data and the extracted college major calculated based on the college code and the major code. The embodiment of the present invention can speed up the matching speed of the resume matching unit by calculating the hash value to be matched and matching using the hash value to be matched.

[0151] Furthermore, the resume matching unit is constructed, including:

[0152] Get the matching company set, extract matching companies from the matching company set in sequence, and perform the following operations on the extracted matching companies:

[0153] The target professional set of the extracted matching companies is obtained, target professional sets are extracted in sequence, and the following operations are performed on the extracted target professional sets:

[0154] Obtain a target school set corresponding to the extracted target major, extract target schools from the target school set in sequence, and perform the following operations on the extracted target schools:

[0155] Constructing unit template division rules based on the extracted target colleges, the extracted target majors and the extracted matching companies, and constructing a recruitment identification code according to the unit template division rules;

[0156] The recruitment identification codes are aggregated to obtain an information classification library, and the resume matching unit is constructed using the information classification library.

[0157] It should be noted that the matching company is the code of the company that expects to recruit, the target major set is the target major that the matching company expects to recruit, and the target school is the school where the target major is offered. The unit template division rule is the rule for classifying resume images. The recruitment identification code is the identification code used to represent the target major and target school that the matching company expects to recruit. The information classification library is a database constructed by multiple recruitment identification codes.

[0158] Exemplarily, the code corresponding to the matching company is 111, the matching company has only one target major and the code corresponding to the target major is 113, and the target major corresponds to two target colleges, 344 and 366 respectively, so the recruitment identification codes are 111-113-344 and 111-113-366. 111-113-344 and 111-113-366 are constructed as two independent unit template division rules, and these two independent unit template division rules are data in the information classification library. The corresponding college code and major code in the resume image are 113 and 344 respectively, so the resume information data is identified as 113-344, so the resume information data is identified as 113-344 and the recruitment identification code is 111-113-344, and the resume matching unit sends the resume image to the matching company.

[0159] S7. After confirming that the matching company set is not an empty set, matching companies are extracted from the matching company set in sequence, and the following operations are performed on the extracted matching companies: obtaining a receiving unit of the extracted matching company, sending a resume image to the receiving unit, and completing the extraction of resume key information of the resume image after the receiving unit confirms receipt of the resume image.

[0160] It is understandable that the receiving unit is a unit that matches the company to receive the resume image. Before the resume image is sent to the receiving unit, the key information of the resume image has been identified. After the resume image is sent to the receiving unit, the resume information data carried by the resume image will be sent together, thereby completing the extraction of the resume key information of the resume image.

[0161] The present invention is to solve the problem described in the background technology. The present invention receives a resume image, performs grayscale processing and correction operations on the resume image, and obtains an initial image. The present invention performs grayscale operations on the resume image to facilitate correction processing. When the characters in the resume image are skewed and distorted, the accuracy of text recognition will be affected. The purpose of correction is to more easily identify the information in the resume image, so the present invention corrects the resume image line by line. Based on a pre-constructed resume template and an initial image, a coordinate set of identification points of an initial image is obtained; based on the initial image, a coordinate set to be filled in is extracted from the coordinate set of identification points in sequence; and the following operations are performed on the extracted coordinate sets to be filled in: a window to be filled corresponding to the extracted coordinate set to be filled is obtained, and a keyword identifier of the window to be filled is confirmed; the pixel sum of the extracted window to be filled is obtained; after confirming that the window to be filled has been filled based on the pixel sum, a pre-constructed text segmentation method is used to divide the window to be filled to obtain a keyword image sequence; a pre-constructed text recognition method is used to recognize the keyword image sequence to obtain key text; a marking operation is performed on the key text using the keyword identifier to obtain unit key data; the present invention detects each window to be filled in. Since the initial image has been obtained, compared with directly extracting text and finding that the window to be filled has not been filled in, and then directly using the pixel sum to judge whether it is filled in, it can judge whether the window to be filled in is filled in more quickly. Summarize the key data of the unit, obtain the resume information data corresponding to the resume image, obtain the matching company set of the resume image based on the resume information data, confirm that the matching company set is not an empty set, extract the matching companies from the matching company set in sequence, and perform the following operations on the extracted matching companies: obtain the receiving unit of the extracted matching company, send the resume image to the receiving unit, and complete the extraction of the resume key information of the resume image after the receiving unit confirms the receipt of the resume image. The present invention extracts the college and major from the resume information, and sends the patent image to the resume matching unit according to the college and major, uses the data in the resume matching unit to match the major and college corresponding to the resume image with the college and major expected by the matching company, and finally sends the resume image to each matching company in the matching company set, that is, the embodiment of the present invention can extract the major and college from the resume image, and send the resume image that meets the recruitment expectations of the matching company to the receiving unit of the matching company, which is convenient for job seekers to send resumes and also conducive to matching companies to receive resumes. Therefore, the present invention extracts key information from the resume image and sends the resume to the matching company that meets the conditions according to the key information.

[0162] like Figure 2 , which is a functional module diagram of a resume key information extraction system provided by an embodiment of the present invention.

[0163] The resume key information extraction system 100 of the present invention can be installed in an electronic device. According to the functions to be implemented, the resume key information extraction system 100 can include an image preprocessing module 101, a unit key data acquisition module 102, a matching company acquisition module 103 and an information sending module 104. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0164] The image preprocessing module 101 is used to receive a resume image, perform grayscale processing and correction operations on the resume image, and obtain an initial image;

[0165] The unit key data acquisition module 102 is used to acquire a set of identification point coordinates of an initial image based on a pre-constructed resume template and an initial image, wherein the set of identification point coordinates includes a plurality of to-be-filled coordinate sets, and the to-be-filled coordinate sets are sequentially extracted from the set of identification point coordinates based on the initial image, and the following operations are performed on the extracted to-be-filled coordinate sets: obtaining a to-be-filled window corresponding to the extracted to-be-filled coordinate set, and confirming a keyword identifier of the to-be-filled window, obtaining a pixel sum of the extracted to-be-filled window, confirming that the to-be-filled window has been filled based on the pixel sum, dividing the to-be-filled window by a pre-constructed text segmentation method, obtaining a keyword image sequence, recognizing the keyword image sequence by a pre-constructed text recognition method, obtaining key text, and using the keyword identifier to identify the keyword image sequence. Performing a marking operation on the key text to obtain unit key data, wherein the obtaining of the pixel sum of the extracted window to be filled, and confirming that the window to be filled has been filled based on the pixel sum, comprises: performing a binarization operation on the window to be filled using a pre-built binarization method to obtain a binary window image, obtaining multiple pixel values ​​of the binary window image, calculating a pixel sum according to the multiple pixel values, comparing the pixel sum with a preset sum threshold, if the pixel sum is less than or equal to the sum threshold, confirming that the window to be filled has not been filled, performing a not-filled marking operation on the window to be filled to obtain an not-filled marking window, confirming that the window to be filled has been filled based on the not-filled marking window, and if the pixel sum is greater than the sum threshold, confirming that the window to be filled has been filled;

[0166] The matching company acquisition module 103 is used to summarize the key data of the unit, obtain the resume information data corresponding to the resume image, and obtain the matching company set of the resume image based on the resume information data, wherein the acquisition of the matching company set of the resume image based on the resume information data includes: constructing a resume matching unit, extracting school data and major data based on the resume information data, obtaining the school code and major code based on the school data and major data, calculating the hash value to be matched according to the school code and major code, inputting the hash value to be matched into the resume matching unit, and obtaining the matching company set, wherein the construction of the resume matching unit includes: obtaining the matching company set, extracting the school data and major data based on the resume information data, obtaining the school code and major code based on the school data and major data, calculating the hash value to be matched according to the school code and major code, and inputting the hash value to be matched into the resume matching unit to obtain the matching company set. Extract matching companies in sequence, and perform the following operations on the extracted matching companies: obtain the target major set of the extracted matching companies, extract target majors in sequence from the target major set, and perform the following operations on the extracted target majors: obtain the target school set corresponding to the extracted target major, extract target schools in sequence from the target school set, and perform the following operations on the extracted target schools: construct unit template division rules based on the extracted target schools, the extracted target majors and the extracted matching companies, construct recruitment identification codes according to the unit template division rules, summarize the recruitment identification codes, obtain an information classification library, and use the information classification library to construct a resume matching unit;

[0167] The information sending module 104 is used to confirm that the matching company set is not an empty set, extract matching companies from the matching company set in sequence, and perform the following operations on the extracted matching companies: obtain the receiving unit of the extracted matching company, send the resume image to the receiving unit, and complete the extraction of resume key information of the resume image after the receiving unit confirms the receipt of the resume image.

[0168] In detail, each module in the resume key information extraction system 100 in the embodiment of the present invention is used in the same manner as described above. Figure 1 The same technical means are used as the resume key information extraction method described in , and can produce the same technical effects, so I will not go into details here.

[0169] like Figure 3 , is a schematic diagram of the structure of an electronic device for implementing a resume key information extraction method provided by an embodiment of the present invention.

[0170] The electronic device 1 may include a processor 10, a memory 11 and a bus 12, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a resume key information extraction method program.

[0171] The memory 11 includes at least one type of readable storage medium, and the readable storage medium includes a flash memory, a mobile hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory, etc.), a magnetic memory, a disk, an optical disk, etc. In some embodiments, the memory 11 may be an internal storage unit of the electronic device 1, such as a mobile hard disk of the electronic device 1. In other embodiments, the memory 11 may also be an external storage device of the electronic device 1, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device 1. Further, the memory 11 also includes an internal storage unit of the electronic device 1 and an external storage device. The memory 11 can not only be used to store application software and various types of data installed in the electronic device 1, such as the code of the resume key information extraction method program, etc., but also can be used to temporarily store data that has been output or is to be output.

[0172] The processor 10 may be composed of an integrated circuit in some embodiments, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips, etc. The processor 10 is the control core (Control Unit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and executes or executes programs or modules (such as resume key information extraction method programs, etc.) stored in the memory 11, and calls data stored in the memory 11 to execute various functions of the electronic device 1 and process data.

[0173] The bus 12 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 may be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize connection and communication between the memory 11 and at least one processor 10, etc.

[0174] Figure 3 Only an electronic device with components is shown, and those skilled in the art will understand that Figure 3The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0175] For example, although not shown, the electronic device 1 may also include a power source (such as a battery) for supplying power to various components. Preferably, the power source may be logically connected to the at least one processor 10 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 1 may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0176] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device 1 and other electronic devices.

[0177] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), or a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode) touch device. The display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device 1 and to display a visual user interface.

[0178] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0179] The resume key information extraction method program stored in the memory 11 of the electronic device 1 is a combination of multiple instructions. When running in the processor 10, it can achieve:

[0180] Receive a resume image, perform grayscale processing and correction operations on the resume image to obtain an initial image, obtain a marker point coordinate set of the initial image based on a pre-built resume template and the initial image, wherein the marker point coordinate set includes a plurality of coordinate sets to be filled in, extract the coordinate sets to be filled in sequence from the marker point coordinate set based on the initial image, and perform the following operations on the extracted coordinate sets to be filled in:

[0181] Obtaining a window to be filled corresponding to the extracted coordinate set to be filled, and confirming a keyword identifier of the window to be filled, obtaining a pixel sum of the extracted window to be filled, and after confirming that the window to be filled has been filled based on the pixel sum, dividing the window to be filled using a pre-built text segmentation method to obtain a keyword image sequence;

[0182] The step of obtaining the pixel sum of the extracted window to be filled, and confirming that the window to be filled has been filled based on the pixel sum, comprises:

[0183] Using a pre-built binarization method to perform a binarization operation on the window to be filled, obtaining a binary window image, obtaining multiple pixel values ​​of the binary window image, and calculating a pixel sum according to the multiple pixel values;

[0184] Comparing the pixel sum with a preset sum threshold;

[0185] If the pixel sum is less than or equal to the sum threshold, it is confirmed that the window to be filled is not filled, and after performing an unfilled marking operation on the window to be filled, an unfilled marking window is obtained, and it is confirmed that the window to be filled is filled based on the unfilled marking window;

[0186] If the pixel sum is greater than the sum threshold, it is confirmed that the window to be filled has been filled;

[0187] Using a pre-built text recognition method to recognize the keyword image sequence to obtain key text, and using a keyword identifier to perform an identification operation on the key text to obtain unit key data;

[0188] Summarize key data of the unit to obtain resume information data corresponding to the resume image, and obtain a matching company set of the resume image based on the resume information data;

[0189] Wherein, the step of obtaining a matching company set of the resume image based on the resume information data includes:

[0190] Constructing a resume matching unit;

[0191] Extracting school data and major data based on the resume information data, obtaining school code and major code based on the school data and major data, calculating a hash value to be matched according to the school code and major code, inputting the hash value to be matched into a resume matching unit, and obtaining a matching company set;

[0192] The construction of the resume matching unit includes:

[0193] Get the matching company set, extract matching companies from the matching company set in sequence, and perform the following operations on the extracted matching companies:

[0194] The target professional set of the extracted matching companies is obtained, target professional sets are extracted in sequence, and the following operations are performed on the extracted target professional sets:

[0195] Obtain a target school set corresponding to the extracted target major, extract target schools from the target school set in sequence, and perform the following operations on the extracted target schools:

[0196] Constructing unit template division rules based on the extracted target colleges, the extracted target majors and the extracted matching companies, and constructing a recruitment identification code according to the unit template division rules;

[0197] Aggregate the recruitment identification codes to obtain an information classification library, and use the information classification library to construct a resume matching unit;

[0198] After confirming that the matching company set is not an empty set, matching companies are extracted from the matching company set in sequence, and the following operations are performed on the extracted matching companies:

[0199] The receiving unit of the extracted matching company is obtained, and the resume image is sent to the receiving unit. After the receiving unit confirms receipt of the resume image, the resume key information of the resume image is extracted.

[0200] Specifically, the specific implementation method of the processor 10 for the above instructions can refer to Figures 1 to 3 The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0201] Furthermore, if the module / unit integrated in the electronic device 1 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

[0202] The present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor of an electronic device, the computer program can implement:

[0203] Receive a resume image, perform grayscale processing and correction operations on the resume image to obtain an initial image, obtain a marker point coordinate set of the initial image based on a pre-built resume template and the initial image, wherein the marker point coordinate set includes a plurality of coordinate sets to be filled in, extract the coordinate sets to be filled in sequence from the marker point coordinate set based on the initial image, and perform the following operations on the extracted coordinate sets to be filled in:

[0204] Obtaining a window to be filled corresponding to the extracted coordinate set to be filled, and confirming a keyword identifier of the window to be filled, obtaining a pixel sum of the extracted window to be filled, and after confirming that the window to be filled has been filled based on the pixel sum, dividing the window to be filled using a pre-built text segmentation method to obtain a keyword image sequence;

[0205] The step of obtaining the pixel sum of the extracted window to be filled, and confirming that the window to be filled has been filled based on the pixel sum, comprises:

[0206] Using a pre-built binarization method to perform a binarization operation on the window to be filled, obtaining a binary window image, obtaining multiple pixel values ​​of the binary window image, and calculating a pixel sum according to the multiple pixel values;

[0207] Comparing the pixel sum with a preset sum threshold;

[0208] If the pixel sum is less than or equal to the sum threshold, it is confirmed that the window to be filled is not filled, and after performing an unfilled marking operation on the window to be filled, an unfilled marking window is obtained, and it is confirmed that the window to be filled is filled based on the unfilled marking window;

[0209] If the pixel sum is greater than the sum threshold, it is confirmed that the window to be filled has been filled;

[0210] Using a pre-built text recognition method to recognize the keyword image sequence to obtain key text, and using a keyword identifier to perform an identification operation on the key text to obtain unit key data;

[0211] Summarize key data of the unit to obtain resume information data corresponding to the resume image, and obtain a matching company set of the resume image based on the resume information data;

[0212] Wherein, the step of obtaining a matching company set of the resume image based on the resume information data includes:

[0213] Constructing a resume matching unit;

[0214] Extracting school data and major data based on the resume information data, obtaining school code and major code based on the school data and major data, calculating a hash value to be matched according to the school code and major code, inputting the hash value to be matched into a resume matching unit, and obtaining a matching company set;

[0215] The construction of the resume matching unit includes:

[0216] Get the matching company set, extract matching companies from the matching company set in sequence, and perform the following operations on the extracted matching companies:

[0217] The target professional set of the extracted matching companies is obtained, target professional sets are extracted in sequence, and the following operations are performed on the extracted target professional sets:

[0218] Obtain a target school set corresponding to the extracted target major, extract target schools from the target school set in sequence, and perform the following operations on the extracted target schools:

[0219] Constructing unit template division rules based on the extracted target colleges, the extracted target majors and the extracted matching companies, and constructing a recruitment identification code according to the unit template division rules;

[0220] Aggregate the recruitment identification codes to obtain an information classification library, and use the information classification library to construct a resume matching unit;

[0221] After confirming that the matching company set is not an empty set, matching companies are extracted from the matching company set in sequence, and the following operations are performed on the extracted matching companies:

[0222] The receiving unit of the extracted matching company is obtained, and the resume image is sent to the receiving unit. After the receiving unit confirms receipt of the resume image, the resume key information of the resume image is extracted.

[0223] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, systems and methods can be implemented in other ways. For example, the system embodiments described above are only illustrative, and actual implementation may have other division methods.

[0224] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0225] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0226] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0227] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claim can also be implemented by one unit or device through software or hardware. The second and other words are used to indicate names, but not to indicate any particular order.

[0228] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for extracting key information from a resume, characterized in that: The method comprises: Receive a resume image, perform grayscale processing and correction operations on the resume image to obtain an initial image, obtain a marker point coordinate set of the initial image based on a pre-built resume template and the initial image, wherein the marker point coordinate set includes a plurality of coordinate sets to be filled in, extract the coordinate sets to be filled in sequence from the marker point coordinate set based on the initial image, and perform the following operations on the extracted coordinate sets to be filled in: Obtaining a window to be filled corresponding to the extracted coordinate set to be filled, and confirming a keyword identifier of the window to be filled, obtaining a pixel sum of the extracted window to be filled, and after confirming that the window to be filled has been filled based on the pixel sum, dividing the window to be filled using a pre-built text segmentation method to obtain a keyword image sequence; The step of obtaining the pixel sum of the extracted window to be filled, and confirming that the window to be filled has been filled based on the pixel sum, comprises: Using a pre-built binarization method to perform a binarization operation on the window to be filled, obtaining a binary window image, obtaining multiple pixel values ​​of the binary window image, and calculating a pixel sum according to the multiple pixel values; Comparing the pixel sum with a preset sum threshold; If the pixel sum is less than or equal to the sum threshold, it is confirmed that the window to be filled is not filled, and after performing an unfilled marking operation on the window to be filled, an unfilled marking window is obtained, and it is confirmed that the window to be filled is filled based on the unfilled marking window; If the pixel sum is greater than the sum threshold, it is confirmed that the window to be filled has been filled; Using a pre-built text recognition method to recognize the keyword image sequence to obtain key text, and using a keyword identifier to perform an identification operation on the key text to obtain unit key data; Summarize key data of the unit to obtain resume information data corresponding to the resume image, and obtain a matching company set of the resume image based on the resume information data; Wherein, the step of obtaining a matching company set of the resume image based on the resume information data includes: Constructing a resume matching unit; Extracting school data and major data based on the resume information data, obtaining school code and major code based on the school data and major data, calculating a hash value to be matched according to the school code and major code, inputting the hash value to be matched into a resume matching unit, and obtaining a matching company set; The construction of the resume matching unit includes: Get the matching company set, extract matching companies from the matching company set in sequence, and perform the following operations on the extracted matching companies: The target professional set of the extracted matching companies is obtained, target professional sets are extracted in sequence, and the following operations are performed on the extracted target professional sets: Obtain a target school set corresponding to the extracted target major, extract target schools from the target school set in sequence, and perform the following operations on the extracted target schools: Constructing unit template division rules based on the extracted target colleges, the extracted target majors and the extracted matching companies, and constructing a recruitment identification code according to the unit template division rules; Aggregate the recruitment identification codes to obtain an information classification library, and use the information classification library to construct a resume matching unit; After confirming that the matching company set is not an empty set, matching companies are extracted from the matching company set in sequence, and the following operations are performed on the extracted matching companies: The receiving unit of the extracted matching company is obtained, and the resume image is sent to the receiving unit. After the receiving unit confirms receipt of the resume image, the resume key information of the resume image is extracted.

2. The method for extracting key information from a resume as claimed in claim 1, characterized in that: The grayscale processing and correction operation is performed on the resume image to obtain the initial image, including: Acquire a plurality of pixels based on the resume image, extract pixels from the plurality of pixels in sequence, and perform the following operations on the extracted pixels: Obtaining the red channel value, green channel value, and blue channel value of the extracted pixel, inputting the red channel value, green channel value, and blue channel value into a pre-built grayscale conversion formula to obtain a grayscale value, and assigning the grayscale value to the extracted pixel to obtain a grayscale pixel; Aggregate the grayscale pixels to get a grayscale image; A correction operation is performed on the grayscale image to obtain an initial image.

3. The method for extracting key information from a resume as claimed in claim 2, characterized in that: The step of performing a correction operation on the grayscale image to obtain an initial image includes: A grayscale image coordinate system corresponding to the grayscale image is obtained, an image vertical axis of the grayscale image is obtained based on the grayscale image coordinate system, a vertical projection operation is performed on the grayscale image based on the image vertical axis to obtain a plurality of row boundary sets, row boundary sets are sequentially extracted from the plurality of row boundary sets, and the following operations are performed on the extracted row boundary sets: Fitting an upper boundary slope based on the extracted row boundary set, and comparing the upper boundary slope with a preset slope threshold; If the upper boundary slope is greater than the slope threshold, obtaining a line text image corresponding to the upper boundary slope, and performing a correction operation on the line text image to obtain a corrected line image; Otherwise, obtaining a line text image corresponding to the upper boundary slope, and confirming that the line text image is a line image that does not need correction; Summarize the corrected row images to obtain a corrected row image set, and summarize the row images that do not need to be corrected to obtain an original row image set; After constructing a first correction image based on the corrected row image set and the original row image set, obtaining a correction number according to the first correction image; Comparing the correction number with a preset correction number threshold, if the correction number is greater than the correction number threshold, confirming that the first corrected image corresponding to the correction number is the initial image; Otherwise, confirm that the first corrected image is a grayscale image, and return to the step of obtaining the grayscale image coordinates corresponding to the grayscale image, until the corrected row image set is an empty set or the correction number is greater than the correction number threshold, to obtain the initial image; The step of fitting the upper boundary slope based on the extracted row boundary set comprises: Acquire multiple row upper boundary pixels in the row boundary set, extract the pixel horizontal coordinate of each row upper boundary pixel in turn from the multiple row upper boundary pixels to obtain a pixel horizontal coordinate set, perform a linear fitting operation on the pixel horizontal coordinate set to obtain an upper boundary straight line, and calculate the upper boundary slope based on the upper boundary straight line.

4. The method for extracting key information from a resume as claimed in claim 3, characterized in that: The method of dividing the window to be filled by using the pre-built text segmentation method to obtain a keyword image sequence includes: Acquire a binary window image of the window to be filled, acquire a binary coordinate system based on the binary window image, acquire a binary vertical axis based on the binary coordinate system, perform a vertical projection operation on the binary window image according to the binary vertical axis, and obtain a binary row pixel set; Dividing the binary row pixel set to obtain a plurality of unit row pixel sets; Extracting unit row pixel sets from the plurality of unit row pixel sets in sequence, and performing the following operations on the extracted unit row pixel sets: Obtain a binary row image based on the extracted unit row pixel set, obtain a binary horizontal axis based on a binary coordinate system, perform a vertical projection operation on the binary row image based on the binary horizontal axis to obtain a binary vertical pixel set, and divide the binary vertical pixel set using a pre-constructed character division method to obtain a keyword image; Summarize keyword images to obtain keyword image sequences.

5. The method for extracting key information from a resume as claimed in claim 4, characterized in that: The method of using a pre-built binarization method to perform a binarization operation on the window to be filled to obtain a binary window image includes: A grayscale set is obtained based on the window to be filled, grayscale values ​​are extracted from the grayscale set in sequence, and the following operations are performed on the extracted grayscale values: The extracted grayscale value is removed from the grayscale set to obtain a grayscale value set to be classified. After confirming that one or more pixels corresponding to the extracted grayscale value are pixels of the same type, the grayscale values ​​to be classified are sequentially extracted from the grayscale value set to be classified, and the following operations are performed on the extracted grayscale values ​​to be classified: Compare the extracted grayscale value to be classified with the extracted grayscale level value; If the extracted grayscale value to be classified is less than or equal to the extracted grayscale level value, one or more pixels corresponding to the extracted grayscale value to be classified are confirmed as first-class pixels; otherwise, one or more pixels corresponding to the extracted grayscale value to be classified are confirmed as second-class pixels; The first-class pixels and the second-class pixels are respectively summarized to obtain the first-class pixel set and the second-class pixel set, and the first-class pixel probability and the second-class pixel probability are respectively calculated based on the first-class pixel set and the second-class pixel set, wherein the first-class pixel probability is: in, represents the probability of a class of pixels, Indicates the number of grayscale values ​​to be classified. Indicates the number of pixels corresponding to the grayscale value to be divided, Indicates the total number of pixels in the window to be filled; The maximum inter-class variance of the extracted grayscale values ​​is calculated based on the probability of the first class pixels and the probability of the second class pixels, wherein the calculation formula of the maximum inter-class variance is: in, represents the maximum between-class variance, represents the probability of a class of pixels, represents the probability of two types of pixels, Represents the mean gray value of pixels in a class of pixels. Represents the mean grayscale value of all pixels corresponding to the window to be filled, Represents the mean of the grayscale values ​​of the pixels in the second type of pixel set; The maximum inter-class variances are summarized to obtain a maximum inter-class variance set, an optimal partitioning threshold is determined based on the maximum inter-class variance set, and a binarization operation is performed on the window to be filled according to the optimal partitioning threshold to obtain a binary window image.

6. A resume key information extraction system, characterized in that: The system comprises: An image preprocessing module is used to receive a resume image, perform grayscale processing and correction operations on the resume image, and obtain an initial image; The unit key data acquisition module is used to acquire the identification point coordinate set of the initial image based on the pre-built resume template and the initial image, wherein the identification point coordinate set includes a plurality of coordinate sets to be filled in, and the coordinate sets to be filled in are sequentially extracted from the identification point coordinate set based on the initial image, and the following operations are performed on the extracted coordinate sets to be filled in: obtaining the window to be filled corresponding to the extracted coordinate set to be filled in, and confirming the keyword identifier of the window to be filled in, obtaining the pixel sum of the extracted window to be filled in, confirming that the window to be filled in has been filled in based on the pixel sum, dividing the window to be filled in by using the pre-built text segmentation method, obtaining a keyword image sequence, recognizing the keyword image sequence by using the pre-built text recognition method, obtaining the key text, and using the keyword identifier to identify the key text. The key text performs a marking operation to obtain unit key data, wherein the obtaining of the pixel sum of the extracted window to be filled, and after confirming that the window to be filled has been filled based on the pixel sum, comprises: performing a binarization operation on the window to be filled using a pre-built binarization method to obtain a binary window image, obtaining multiple pixel values ​​of the binary window image, calculating the pixel sum according to the multiple pixel values, comparing the pixel sum with a preset sum threshold, if the pixel sum is less than or equal to the sum threshold, confirming that the window to be filled has not been filled, performing a non-filled marking operation on the window to be filled, obtaining an non-filled marking window, confirming that the window to be filled has been filled based on the non-filled marking window, and if the pixel sum is greater than the sum threshold, confirming that the window to be filled has been filled; The matching company acquisition module is used to summarize the key data of the unit, obtain the resume information data corresponding to the resume image, and obtain the matching company set of the resume image based on the resume information data, wherein the obtaining the matching company set of the resume image based on the resume information data includes: constructing a resume matching unit, extracting school data and major data based on the resume information data, obtaining the school code and major code based on the school data and major data, calculating the hash value to be matched according to the school code and major code, inputting the hash value to be matched into the resume matching unit, and obtaining the matching company set, wherein the constructing of the resume matching unit includes: obtaining the matching company set, extracting the matching company set from the matching company set according to the school code and major code, and obtaining the matching company set according to the school code and major code. Extract matching companies one by one, and perform the following operations on the extracted matching companies: obtain the target major set of the extracted matching companies, extract target majors from the target major set in sequence, and perform the following operations on the extracted target majors: obtain the target school set corresponding to the extracted target major, extract target schools from the target school set in sequence, and perform the following operations on the extracted target schools: construct unit template division rules based on the extracted target schools, the extracted target majors and the extracted matching companies, construct recruitment identification codes according to the unit template division rules, summarize the recruitment identification codes, obtain an information classification library, and use the information classification library to construct a resume matching unit; The information sending module is used to confirm that the matching company set is not an empty set, extract matching companies from the matching company set in sequence, and perform the following operations on the extracted matching companies: obtain the receiving unit of the extracted matching company, send the resume image to the receiving unit, and complete the extraction of resume key information of the resume image after the receiving unit confirms the receipt of the resume image.

7. An electronic device, comprising: A memory storing at least one instruction; and A processor executes instructions stored in the memory to implement the resume key information extraction method described in any one of claims 1-5.

Images