AI-driven graffiti painting emotional psychological state analysis and recognition data processing method and system

Abstract

The invention provides an AI-driven graffiti painting emotional psychological state analysis and recognition data processing method and system, and the method comprises the steps: calling a graffiti recording video corresponding to a target identity inputted by a management end, determining a plurality of key frames corresponding to the graffiti recording video according to a feature analysis model, obtaining the painting elements of the key frames, and carrying out the recognition of the painting elements of the key frames; dividing the key frames according to the drawing elements to obtain a plurality of doodle element groups, constructing a doodle time axis corresponding to the doodle recording video, and generating a plurality of doodle recording nodes corresponding to the doodle element groups on the doodle time axis to obtain a doodle recording axis, and inputting the graffiti recording axis into a graffiti analysis model to obtain graffiti analysis data.

Description

Technical Field

[0001] The present invention relates to data processing technology, and in particular to an AI-driven graffiti painting emotional psychological state analysis and recognition data processing method and system. Background Art

[0002] Graffiti is a magical language that can often be used to express the painter's emotions. It is very helpful for patients, especially children, to adjust their emotions and express their inner thoughts. By analyzing graffiti images, we can provide reference for related research.

[0003] In the prior art, when analyzing graffiti images, the already completed image data is usually analyzed, resulting in relatively single data during image analysis and incomplete analysis data, so that the analyzed data cannot better fit the user.

[0004] Therefore, how to obtain more comprehensive graffiti data and improve the comprehensiveness of graffiti data analysis has become a problem that needs to be solved urgently. Summary of the invention

[0005] The present invention provides an AI-driven graffiti painting emotional psychological state analysis and recognition data processing method and device, which can obtain more comprehensive graffiti data and improve the comprehensiveness of graffiti data analysis.

[0006] A first aspect of the present invention provides an AI-driven graffiti painting emotional psychological state analysis and recognition data processing method, comprising: Retrieving a graffiti recording video corresponding to a target identity input by a management terminal, and determining a plurality of key frames corresponding to the graffiti recording video according to a feature analysis model; Acquire the painting elements of each of the key frames, and divide each of the key frames according to the painting elements to obtain a plurality of graffiti element groups; Constructing a graffiti timeline corresponding to the graffiti recording video, and generating a plurality of graffiti recording nodes corresponding to each of the graffiti element groups on the graffiti timeline to obtain a graffiti recording axis; The graffiti record axis is input into a graffiti analysis model to obtain graffiti analysis data.

[0007] Optionally, in a possible implementation manner of the first aspect, retrieving a graffiti recording video corresponding to a target identity input by a management end, and determining a plurality of key frames corresponding to the graffiti recording video according to a feature analysis model includes: Arranging the recorded frames in the graffiti recorded video in chronological order to obtain a screening sequence, and sequentially acquiring the painting contours in the adjacent recorded frames in the screening sequence based on the feature analysis model; Acquire multiple drawing points corresponding to the drawing outline, perform coordinate processing on the recording frame, and obtain drawing coordinates of each drawing point; Graffiti feature recognition is performed on the recording frame according to the drawing coordinates and the drawing pixel values ​​corresponding to each drawing point, and a recording frame satisfying a graffiti feature condition is determined as a key frame, wherein the graffiti feature condition includes a shape feature condition and a coloring feature condition.

[0008] Optionally, in a possible implementation manner of the first aspect, graffiti feature recognition is performed on the recorded frame according to the drawing coordinates and the drawing pixel values ​​corresponding to each of the drawing points, and a recorded frame satisfying a graffiti feature condition is determined as a key frame, wherein the graffiti feature condition includes a shape feature condition and a coloring feature condition, including: Determine that a drawing point in each recording frame whose drawing coordinates are consistent with those in a previous recording frame is an existing drawing point, and a drawing point in each recording frame whose drawing coordinates are inconsistent with those in a previous recording frame is a newly added drawing point; When the newly added drawing point does not exist in the recording frames of the preset number of consecutive frames, determining that the last recording frame of the consecutive multiple recording frames meets the shape feature condition, and obtaining the corresponding recording frame as a key frame; Obtaining the interval distance between the closest existing drawing point and the newly added drawing point in each of the recording frames; When the interval distance is greater than or equal to the shape filling threshold and / or the drawing pixel value of the newly added drawing point is inconsistent with the existing drawing point, it is determined that the previous recording frame adjacent to the recording frame meets the color filling feature condition, the corresponding recording frame is obtained as a key frame, and the drawing pixel value corresponding to the existing drawing point in the key frame is determined to be the filling color corresponding to the key frame.

[0009] Optionally, in a possible implementation of the first aspect, a plurality of key frames satisfying the shape feature condition are arranged in time order to obtain a shape screening sequence; Sequentially acquiring adjacent key frames in the shape screening sequence as target frames, and extracting target contours in the target frames; Determine that the drawing points in each target outline where the drawing coordinates and the drawing pixel values ​​are consistent are the same drawing points, and count the same number of the same drawing points in the previous target frame; The matching similarity is obtained according to the ratio between the same number and the total number of all drawing points in the previous target frame. When the matching similarity is greater than a similarity threshold, the previous target frame is deleted.

[0010] Optionally, in a possible implementation manner of the first aspect, obtaining a painting element of each key frame, and dividing each key frame according to the painting element to obtain a plurality of graffiti element groups includes: Performing element feature recognition on the painting outline in each of the key frames according to preset element features, and determining the painting element corresponding to each of the key frames; The key frames with the same painting elements are obtained as the same graffiti element group, and the key frames in the graffiti element group that meet the shape feature condition are determined as shape display frames, and the key frames that meet the coloring feature condition are determined as coloring display frames.

[0011] Optionally, in a possible implementation manner of the first aspect, hand contour recognition is performed on each of the recorded frames, and a recorded frame without the hand contour is determined as a pending frame; Arranging the pending frames in reverse time order to obtain a backtracking sequence, and sequentially identifying elements of the pending frames in the backtracking sequence; When the pending frame has drawing elements corresponding to all graffiti element groups, and the clarity of the pending frame is greater than or equal to a clarity threshold, determining that the pending frame is a traceback frame; A graffiti element group whose shape display frame does not meet the shape display condition is obtained as a restricted element group, a lossless frame corresponding to the restricted element group is generated according to the backtracking frame, and the restricted element group is updated based on the lossless frame.

[0012] Optionally, in a possible implementation manner of the first aspect, acquiring a graffiti element group whose shape display frame does not satisfy a shape display condition as a restricted element group, generating a lossless frame corresponding to the restricted element group according to the backtracking frame, and updating the restricted element group based on the lossless frame, including: When the shape display frame has the hand outline and / or the clarity of the shape display frame is less than the clarity threshold, determining that the shape display frame does not meet the shape display condition, and determining that the graffiti element group corresponding to the shape display frame is a restricted element group; Acquire a drawing contour corresponding to the drawing element of the restricted element group in the traceback frame as a lossless contour; The preset lossless pixel value is retrieved to update the image area except the lossless contour in the traceback frame to obtain the lossless frame corresponding to the restricted element group, and the shape display frame in the restricted element group is replaced and updated according to the lossless frame.

[0013] Optionally, in a possible implementation manner of the first aspect, constructing a graffiti timeline corresponding to the graffiti recording video, and generating a plurality of graffiti record nodes corresponding to each of the graffiti element groups on the graffiti timeline to obtain a graffiti record axis includes: Constructing a corresponding graffiti timeline according to the recording time corresponding to the graffiti recording video, obtaining the recording time corresponding to each key frame in each graffiti element group, and determining the graffiti time point corresponding to each recording time in the graffiti timeline; Constructing an initial recording node of a shape display type at a graffiti time point corresponding to the shape display frame, and constructing an initial recording node of a coloring display type at a graffiti time point corresponding to the coloring display frame; Performing shape update on the corresponding initial record node according to the shape of the drawing outline in the shape display frame to obtain a graffiti record node corresponding to the shape display frame; The color of the corresponding initial record node is updated based on the filling color corresponding to the coloring display frame to obtain the graffiti record node corresponding to the coloring display frame, and the graffiti timeline is updated according to the graffiti record node to obtain the graffiti record axis.

[0014] Optionally, in a possible implementation of the first aspect, inputting the graffiti record axis into a graffiti analysis model to obtain graffiti analysis data includes: Retrieving preset graffiti features based on the graffiti analysis model, performing feature comparison on a plurality of graffiti record nodes in the graffiti record axis according to the preset graffiti features, and determining graffiti features corresponding to each of the graffiti record nodes; Wherein, the preset graffiti features include preset shape features and preset color features, and each of the preset graffiti features includes a corresponding preset graffiti icon; The graffiti record axis is updated according to the preset graffiti icons corresponding to the graffiti features to obtain graffiti analysis data corresponding to the target identity.

[0015] A second aspect of the present invention provides an AI-driven graffiti painting emotional psychological state analysis and recognition data processing system, comprising: A calling module, used to call a graffiti recording video corresponding to a target identity input by a management terminal, and determine a plurality of key frames corresponding to the graffiti recording video according to a feature analysis model; An acquisition module, used for acquiring the painting elements of each key frame, and dividing each key frame according to the painting elements to obtain a plurality of graffiti element groups; A construction module, used to construct a graffiti timeline corresponding to the graffiti recording video, and generate a plurality of graffiti record nodes corresponding to each of the graffiti element groups on the graffiti timeline to obtain a graffiti record axis; The analysis module is used to input the graffiti record axis into a graffiti analysis model to obtain graffiti analysis data.

[0016] The beneficial effects of the present invention are as follows: 1. The present invention can record the graffiti process by video, and obtain multiple image frames in the graffiti process through the recorded video data, so that a more comprehensive graffiti data analysis of the user's graffiti process can be performed through image frames at different times, thereby improving the comprehensiveness of data analysis. When analyzing the user's graffiti data, the present invention can remove graffiti image data with consistent features, thereby improving the efficiency of data analysis.

[0017] 2. When analyzing the graffiti shape in the image frame, the present invention can obtain the graffiti shape data and graffiti coloring data at different times, and analyze in combination with the data at different times, so that the data during analysis can be more comprehensive. When acquiring the graffiti shape data, the present invention can first determine the frame without hand contour recording as the pending frame, arrange all the pending frames in reverse chronological order to form a retrospective sequence, and perform element recognition on the pending frames in the retrospective sequence, and determine the pending frames with painting elements corresponding to all graffiti element groups and higher clarity as retrospective frames. By calling the preset lossless pixel value to update the retrospective frame, the corresponding lossless frame containing a complete painting contour and higher clarity can be obtained. According to the lossless frame, the shape display frame that does not meet the shape display condition is replaced and updated, which can ensure that the painting elements are displayed in a more complete and clearer form, thereby improving the accuracy of data analysis.

[0018] 3. The present invention can analyze graffiti data at different times, and can intuitively display the analysis data at different times through the graffiti record axis. When displaying graffiti data, the present invention can display the painter's emotional changes at different times through graffiti icons, thereby providing data reference for managers. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a flowchart of an AI-driven graffiti painting emotional psychological state analysis and recognition data processing method provided by an embodiment of the present invention; Figure 2 is a schematic diagram of a shape screening sequence provided by an embodiment of the present invention; Figure 3 is a schematic diagram of a graffiti recording axis provided by an embodiment of the present invention; Figure 4 It is a structural schematic diagram of an AI-driven graffiti painting emotional psychological state analysis and recognition data processing system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0022] See also Figure 1 , is a schematic diagram of an AI-driven graffiti painting emotional psychological state analysis and recognition data processing method provided by an embodiment of the present invention, Figure 1 The execution subject of the method shown may be a software and / or hardware device. The execution subject of the present application may include but is not limited to at least one of the following: user equipment, network equipment, etc. Among them, user equipment may include but is not limited to computers, smart phones, personal digital assistants (PDA) and the electronic devices mentioned above. Network equipment may include but is not limited to a single network server, a server group consisting of multiple network servers, or a cloud based on cloud computing consisting of a large number of computers or network servers, wherein cloud computing is a type of distributed computing, a super virtual computer composed of a group of loosely coupled computers. This embodiment does not limit this. It includes steps S1 to S4, as follows: S1, retrieve a graffiti recording video corresponding to a target identity input by a management terminal, and determine a plurality of key frames corresponding to the graffiti recording video according to a feature analysis model.

[0023] In order to obtain more comprehensive graffiti data, this solution will record the graffiti process through video, and obtain multiple image frames during the graffiti process through the recorded video data, so that a more comprehensive graffiti data analysis can be performed through the image frames of the user at different times during the graffiti process, thereby improving the comprehensiveness of data analysis.

[0024] It is understandable that by analyzing the image data at different times during the graffiti process, more comprehensive display data can be obtained. Specifically, this solution will combine the analysis results of the image frames to obtain the reference data corresponding to the user's emotions at different times, thereby providing managers with more comprehensive data references.

[0025] Specifically, when a painter is graffitiing, the graffiti process can be recorded by a recording device, for example, the graffiti process can be recorded by a camera to obtain a recorded video of the painter's graffiti, and then the recorded video of the graffiti is bound to the identity information of the painter. When analyzing the painter's graffiti data, the management personnel can input the identity information of the painter on the management end, so as to retrieve the recorded video of the graffiti corresponding to the identity information of the painter, and perform feature recognition on the shape and color of the graffiti in the recorded video of the graffiti according to the feature analysis model, so as to obtain multiple key frames corresponding to the recorded video of the graffiti.

[0026] Among them, the management end refers to the terminal held by the management personnel, such as a computer, the target identity refers to the identity information corresponding to the painter of the graffiti recording video, the graffiti recording video refers to the video of the painter's graffiti process recorded by a recording device, the feature analysis model refers to the model for feature analysis of the image data in the recorded video. Specifically, this solution will identify the shape features and coloring features in the graffiti process, and the key frame refers to the recording frame that meets the graffiti feature conditions.

[0027] Based on the above embodiment, the specific implementation of step S1 may be: S11, arranging the recorded frames in the graffiti recorded video in time sequence to obtain a screening sequence, and sequentially acquiring the drawing contours in the adjacent recorded frames in the screening sequence based on the feature analysis model.

[0028] Specifically, the recording device can collect multiple complete and continuous recording frames during the graffiti process through the camera and save them as video files, and these recording frames can be arranged in chronological order during recording to obtain a screening sequence. This solution will subsequently analyze the shape data and color data in the image, so the recording frames corresponding to each shape completion and the recording frames corresponding to coloring completion can be determined as key frames for subsequent analysis. When determining the recording frames corresponding to the shape completion, the shape drawing process during the graffiti process can be determined through the painting outlines in consecutive adjacent recording frames, and the recording frames corresponding to the completion of the corresponding shape drawing can be determined as key frames.

[0029] S12, obtaining a plurality of drawing points corresponding to the drawing outline, performing coordinate processing on the recording frame, and obtaining drawing coordinates of each of the drawing points.

[0030] It is worth mentioning that the shooting position of the recording device is unchanged, and when the recorded frames are processed in coordinates, the coordinate origin of each recorded frame is consistent.

[0031] The drawing points refer to the contour points of the drawing contours. By determining the drawing coordinates of each drawing contour, the contour points that have changed in adjacent recording frames can be obtained, thereby determining whether the shape drawing process is completed.

[0032] S13, performing graffiti feature recognition on the recording frame according to the drawing coordinates and the drawing pixel values ​​corresponding to each of the drawing points, and determining the recording frame satisfying the graffiti feature condition as a key frame, wherein the graffiti feature condition includes a shape feature condition and a coloring feature condition.

[0033] Among them, the painting pixel value refers to the pixel value corresponding to the painting point. When performing graffiti feature recognition, the shape characteristics of the graffiti can be identified through the painting coordinates, and the color characteristics of the graffiti can be identified through the painting pixel value. The shape feature condition refers to the corresponding condition when the shape is drawn, and the color filling feature condition refers to the corresponding condition when the color filling is completed.

[0034] Specifically, when shape features are recognized through drawing coordinates, whether the drawing process is completed can be determined by whether new drawing coordinates appear in adjacent recording frames. If new drawing coordinates appear, it means that the user may still be drawing the corresponding shape. If no new drawing coordinates appear, it means that the user may have stopped drawing the corresponding shape. At this time, the corresponding recording frame can be determined to be a key frame.

[0035] When identifying the color features of graffiti through painting pixel values, it is possible to determine whether the fill color has changed by determining whether the fill color in adjacent recording frames has changed and whether the fill area of ​​the same color has changed, thereby determining that the corresponding recording frame when the corresponding fill color ends is a key frame.

[0036] Based on the above embodiment, the specific implementation of step S13 may be: S131, determining that the drawing points in each recording frame whose drawing coordinates are consistent with the drawing coordinates in the previous recording frame are existing drawing points, and the drawing points in each recording frame whose drawing coordinates are inconsistent with the drawing coordinates in the previous recording frame are newly added drawing points.

[0037] If there are drawing points in the current recording frame with the same drawing coordinates as those in the previous recording frame, it can be said that these drawing points in the current recording frame already existed in the previous recording frame. Therefore, the drawing points in the current recording frame with the same drawing coordinates as those in the previous recording frame can be determined as existing drawing points. If there are drawing points in the current recording frame with different drawing coordinates from those in the previous recording frame, it can be said that these drawing points in the current recording frame did not exist in the previous recording frame and were newly added. Therefore, the drawing points in the current recording frame with different drawing coordinates from those in the previous recording frame can be determined as newly added drawing points, wherein the existing drawing points refer to the drawing points that already existed in the previous recording frame, and the newly added drawing points refer to the newly added drawing points in the current recording frame.

[0038] S132: When the newly added drawing point does not exist in the recording frames of the preset number of consecutive frames, determine that the last recording frame of the consecutive multiple recording frames meets the shape feature condition, and obtain the corresponding recording frame as a key frame.

[0039] It is understandable that when the trajectory of the drawing point remains relatively stable in multiple consecutive recording frames, that is, no new drawing points appear in multiple consecutive recording frames, then it can be considered that the current graffiti outline has been drawn. Therefore, when there are no new drawing points in multiple consecutive recording frames, it can be considered that the last recording frame in the multiple consecutive recording frames meets the shape feature conditions. Since the last recording frame contains all the information of the current graffiti outline, it can be determined as a key frame. The preset number of frames refers to the number of recording frames corresponding to a preset continuous time period. For example, the number of recording frames within 30 consecutive seconds can be set as the preset number of frames. When no new drawing points appear in the recording frames within 30 consecutive seconds, it can be considered that the last recording frame in the recording frames within 30 consecutive seconds meets the shape feature conditions.

[0040] Based on the above steps, this solution also includes the following embodiments: A1, multiple key frames that meet the shape feature conditions are arranged according to time sequence to obtain a shape screening sequence.

[0041] In actual applications, when the painter is scribbling, the recording device records the complete and continuous scribbling process. When intercepting the recording frames that meet the shape feature conditions, the recording frames corresponding to the combined shapes may be intercepted. Figure 2 , is a schematic diagram of a shape screening sequence provided by an embodiment of the present invention, such as Figure 2As shown, key frame 2 is composed of the circle and multiple line segments in key frame 1, and the shape in key frame 2 includes the shape in key frame 1. Similarly, key frame 3 also includes the shapes in key frame 2 and key frame 1, and key frame 4 includes the shapes in key frame 3, key frame 2, and key frame 1. It can be seen that key frame 4 may be the combined shape finally drawn by the painter, and the previous three key frames are shape data corresponding to the process of drawing the combined shape. By analyzing the combined shape, the properties and characteristics of the shape can be analyzed more accurately, thereby improving the accuracy of the analysis.

[0042] Based on the above reasons, when there is a relationship between the shapes corresponding to multiple frames of images, the final combined shape can be analyzed and the images of the previous shapes can be removed, thereby improving the accuracy of data analysis. It is worth mentioning that when removing images, this solution will only filter out images with a relationship, because for images with a relationship, they actually correspond to the final combined shape, so the image data corresponding to the shapes in the previous stage can be filtered out.

[0043] A2, sequentially acquiring adjacent key frames in the shape screening sequence as target frames, and extracting target contours in the target frames.

[0044] It can be understood that in order to determine whether the graffiti shape in the previous key frame belongs to the graffiti shape in the next key frame, adjacent key frames in the shape screening sequence can be obtained in sequence as target frames, and the shape contour in the target frame can be extracted to obtain the corresponding target contour.

[0045] The target frame refers to the adjacent key frames in the filter sequence, and the target contour refers to the drawing contour extracted from the target frame.

[0046] A3, determining that the drawing points in each target outline whose drawing coordinates and drawing pixel values ​​are consistent are the same drawing points, and counting the same number of the same drawing points in the previous target frame.

[0047] It can be understood that when the shape in the previous key frame belongs to the shape in the next key frame, there will be overlapping parts between the target contour in the previous key frame and the target contour in the next key frame, and the drawing coordinates and drawing pixel values ​​corresponding to the drawing points of the target contour in the overlapping part will also be consistent. Therefore, the drawing coordinates and drawing pixel values ​​corresponding to the target contours in adjacent target frames can be compared. According to the comparison results, the drawing points with consistent drawing coordinates and drawing pixel values ​​can be determined as the same drawing points, thereby counting the number of identical drawing points in the previous target frame.

[0048] For example Figure 2As shown, when the shape in key frame 2 contains the shape in key frame 1, the circular part in key frame 1 will overlap with the circular part in key frame 2, and the contour points corresponding to the circular part in key frame 1 will be consistent with the contour points corresponding to the circular part in key frame 2, that is, the contour points of the circular part in key frame 1 will be determined as the same drawing points. By counting the number of the same drawing points and the ratio between the subsequent and all contour points, it can be determined whether the shape in the previous frame image belongs to the next frame image.

[0049] A4, obtaining a matching similarity according to a ratio between the same number and the total number of all drawing points in the previous target frame, and deleting the previous target frame when the matching similarity is greater than a similarity threshold.

[0050] Among them, matching similarity refers to a value that can be used to determine the relationship between two contours, and the similarity threshold is a threshold used to determine the relationship between shapes. If it is greater than the threshold, it means that the previous shape may be a component of the next shape. If it is less than or equal to the threshold, it means that the previous shape may not be a component of the next shape. In practical applications, the similarity threshold can be a value close to 100%, such as 98%.

[0051] For example Figure 2 As shown, when the shape in key frame 1 belongs to the shape in key frame 2, the ratio obtained by the same number and the total number may be close to 100%. This is because when the shape in key frame 1 belongs to the shape in key frame 2, the corresponding contour points will be close to 100% consistent with the contour points corresponding to the corresponding part in key frame 2. Therefore, when it is determined that the shapes in adjacent images are in a belonging relationship, the previous image frame can be deleted and only the next image frame is retained.

[0052] Through the above implementation, the accuracy and efficiency of image analysis can be improved.

[0053] S133: Obtain the interval distance between the closest existing drawing point and the newly added drawing point in each of the recording frames.

[0054] It is understandable that when filling the graffiti shapes in the recorded frames with colors, the graffiti shapes may be filled with different colors, or different graffiti shapes may be filled with colors. In order to obtain more comprehensive coloring data, this solution can capture the recorded frames of different coloring situations separately to obtain the coloring data at different times, so that data analysis can be performed on the coloring data at different times.

[0055] Specifically, when the pixel values ​​of the newly added drawing points are inconsistent with those of the existing drawing points, the filling color may change, and the recording frame at the corresponding moment may be intercepted. When the distance between the newly added drawing points and the existing drawing points is too far, the shape of the filling color may change, and the recording frame at the corresponding moment may also be intercepted. The interval distance refers to the distance between the closest existing drawing point and the newly added drawing point.

[0056] S134, when the interval distance is greater than or equal to the shape filling threshold and / or the drawing pixel value of the newly added drawing point is inconsistent with the existing drawing point, determine that the previous recording frame adjacent to the recording frame meets the color filling feature condition, obtain the corresponding recording frame as a key frame, and determine that the drawing pixel value corresponding to the existing drawing point in the key frame is the filling color corresponding to the key frame.

[0057] In actual applications, since the analysis data corresponding to different coloring may be different, and the coloring analysis data of different shapes may also be different, when acquiring the coloring data during the graffiti process, the coloring data when filling with different colors can be acquired, or the corresponding data can be acquired in combination with the graffiti shape corresponding to the filling color.

[0058] Specifically, when the interval distance is greater than the shape filling threshold, it can be considered that the distance between the newly added drawing point and the existing drawing point is too far. At this time, it can be considered that the painter may have changed the graffiti shape corresponding to the filling color, indicating that the color of the corresponding graffiti shape in the previous recording frame has been filled. Therefore, it can be determined that the previous recording frame adjacent to the current recording frame meets the color filling feature conditions, and the corresponding recording frame is determined as a key frame.

[0059] When the drawing pixel value of the newly added drawing point is inconsistent with the drawing pixel value of the existing drawing point, it can be considered that the filling color has changed. At this time, the previous filling color has been filled in the previous recording frame. Therefore, the recording frame corresponding to the moment when the previous color is filled can be intercepted. Specifically, it can be determined that the previous recording frame adjacent to the current recording frame meets the color filling feature conditions, and the corresponding recording frame is determined as a key frame.

[0060] Since this solution aims to obtain the data when the filling color is completed, when the filling color changes, it can be considered that the corresponding filling color in the previous frame has been filled. Therefore, the corresponding filling data of the previous recording frame can be obtained. In order to determine the filling color corresponding to the key frame, the painting pixel value corresponding to the existing drawing point in the key frame can be extracted, and the painting pixel value corresponding to the existing drawing point is determined as the filling color of the key frame.

[0061] Specifically, when there is only one pixel value in the key frame, the corresponding pixel value can be used as the filling color. When there are multiple pixel values ​​in the key frame, the current filling color can be determined by the determination time of each pixel value. For example, when there are pixel values ​​corresponding to the three colors of red, yellow and blue in the key frame, the filling color corresponding to the current image frame can be determined by determining the historical determination time of each pixel value. Since each time the pixel value changes, a corresponding record is made, the recording time of each pixel value in the key frame can be obtained, so that the only pixel value without a recording time can be used as the filling color corresponding to the current image frame. For example, the first filling is red. When the filling color changes to blue, red will be recorded and the corresponding key frame will be obtained. When the blue filling is completed and converted to yellow, there will be two pixel values ​​of blue and red in the previous frame. Therefore, red has been recorded before, and the filling color corresponding to the previous frame is blue. Therefore, the key frame corresponding to each filling color can be obtained.

[0062] The shape filling threshold refers to a preset distance value used to determine whether to fill the same graffiti shape with color.

[0063] Through the above implementation, the graffiti shape data and coloring data within the corresponding time period can be acquired, so that more comprehensive analysis data can be obtained.

[0064] S2, obtaining the painting elements of each of the key frames, and dividing each of the key frames according to the painting elements to obtain a plurality of graffiti element groups.

[0065] In practical applications, painting elements can be composed of shape elements and color elements. The shape elements have corresponding shape analysis data, and the color elements have corresponding color analysis data. After the shape elements and color elements corresponding to the same painting element are classified into the same graffiti element group, the shape data and color data corresponding to the graffiti element group can be subsequently analyzed accordingly.

[0066] Among them, the painting elements refer to the graffiti objects in the key frames, such as the sun, flowers, etc., and the graffiti element group refers to the element group composed of the key frames corresponding to the shape elements that constitute the complete painting elements and the key frames corresponding to the color elements. The graffiti element group can be used to perform specific data analysis on the painting elements.

[0067] In some embodiments, step S2 includes S21 to S22, which are specifically as follows: S21, performing element feature recognition on the painting outline in each of the key frames according to preset element features, and determining the painting element corresponding to each of the key frames.

[0068] Specifically, when determining the painting element corresponding to the key frame, key features such as shape, color, etc. can be extracted from the painting outline in the key frame, and the extracted key features and pre-set element features are used for feature recognition to determine the painting element corresponding to each key frame. The preset element feature refers to the feature corresponding to the preset painting element, which can be used to identify the identity of the graffiti object.

[0069] S22, acquiring key frames with the same painting elements as the same graffiti element group, determining the key frames in the graffiti element group that meet the shape feature condition as shape display frames, and the key frames that meet the coloring feature condition as coloring display frames.

[0070] After identifying the painting elements of each key frame, since the same painting element may correspond to multiple key features, a painting element will correspond to multiple key frames. The key frames with the same painting elements are grouped together, namely, a graffiti element group. Subsequently, different data analyses can be performed on the key frames in the graffiti element group according to different features. Specifically, the key frames that meet the shape feature conditions can be marked as shape display frames to display the shape features of the painting elements, and the key frames that meet the coloring feature conditions can be marked as coloring display frames to display the coloring effects of the painting elements.

[0071] Through the above implementation, it is convenient to subsequently generate corresponding record nodes, and the graffiti data can be better displayed.

[0072] Based on the above steps, this solution also includes the following embodiments: B1, performing hand contour recognition on each of the recorded frames, and determining the recorded frames without the hand contour as pending frames.

[0073] It is understandable that during the graffiti process, the graffiti data may be blocked by the painter's hand outline. In order to improve the accuracy of subsequent data analysis, the recording frame with hand outline blockage can be replaced so that the recording frame during analysis contains complete graffiti data, which can improve the accuracy of data analysis.

[0074] Specifically, the contour extraction technology in the prior art can be used to identify whether there is a hand contour of the painter in each recording frame, and the recording frames without hand contours are determined as pending frames, and the pending frames can be further screened later.

[0075] B2, arranging the pending frames in reverse time order to obtain a backtracking sequence, and performing element identification on the pending frames in the backtracking sequence in sequence.

[0076] It is understandable that when the painter completes a complete graffiti shape, the degree of completion of the graffiti shape in the later recorded frames is higher than that in the previous ones. Therefore, when starting to recognize from the later recorded frames, the recorded frames with higher completion can be analyzed first, so that the recorded frames containing all the painting elements can be quickly identified, which can improve the efficiency of data processing.

[0077] Specifically, all pending frames are arranged in reverse chronological order, that is, arranged from back to front in chronological order to form a backtracking sequence. By performing element recognition on the pending frames in the backtracking sequence, element recognition can be performed on the pending frames in sequence starting from the most recent pending frame without a hand contour, and backtracking to the pending frame containing all painting elements, wherein the backtracking sequence refers to a sequence obtained by arranging the pending frames in reverse chronological order.

[0078] B3: When the pending frame contains drawing elements corresponding to all graffiti element groups and the clarity of the pending frame is greater than or equal to a clarity threshold, determine that the pending frame is a traceback frame.

[0079] The clarity threshold refers to a preset value for measuring the clarity of the pending frame. The larger the clarity threshold, the higher the clarity. In order to improve the accuracy of extracting graffiti shapes, it is necessary to ensure that the pending frame has a high clarity. A relatively clear pending frame can be obtained through the preset clarity threshold. The retrospective frame refers to a pending frame that contains painting elements corresponding to all graffiti element groups and whose clarity is greater than or equal to the clarity threshold.

[0080] B4, acquiring a graffiti element group whose shape display frame does not meet the shape display condition as a restricted element group, generating a lossless frame corresponding to the restricted element group according to the backtracking frame, and updating the restricted element group based on the lossless frame.

[0081] Since this solution will extract and analyze the shapes when analyzing the shapes, if it is found that the shape display frames corresponding to some graffiti element groups do not meet the shape display conditions, for example, the shape display frames may be blocked by the hand contour or the clarity is not high, etc., it may affect the accuracy of data analysis. In order to improve the accuracy of data analysis, the corresponding graffiti element group can be determined as a restricted element group. According to the backtracking frames determined above, the lossless frames corresponding to the restricted element group can be generated, and the shape display frames that do not meet the shape display conditions in the restricted element group can be replaced with the lossless frames later.

[0082] Among them, lossless frame refers to a frame that can clearly and completely display the graffiti shape, shape display condition refers to the absence of hand contour occlusion and high clarity, and restricted element group refers to a graffiti element group whose shape display frame does not meet the shape display condition.

[0083] Based on the above embodiment, the specific implementation of step B4 may be: B41: When the shape display frame contains the hand outline and / or the clarity of the shape display frame is less than the clarity threshold, it is determined that the shape display frame does not meet the shape display condition, and the graffiti element group corresponding to the shape display frame is determined to be a restricted element group.

[0084] Specifically, it is first determined whether the shape display frame meets the shape display conditions. The shape display conditions include two aspects: first, there is no hand outline in the shape display frame. If it exists, the hand outline may obscure the graffiti shape in the shape display frame. When obtaining the shape data in the shape display frame, the complete shape data may not be obtained. Therefore, when the shape display frame does not meet the shape display conditions, the shape display frame can be replaced. Second, the clarity of the shape display frame is greater than or equal to a preset clarity threshold. If it is less than, it indicates that the image quality of the frame is not enough to clearly display the shape of the painting element. When the shape display frame does not meet any of the above conditions, the graffiti element group corresponding to the frame can be determined as a restricted element group, that is, the graffiti element group needs to be subsequently replaced.

[0085] B42, obtaining a drawing contour corresponding to the drawing element of the restricted element group in the traceback frame as a lossless contour.

[0086] Specifically, the retrospective frame contains a variety of complete and clear painting elements, and the painting contour corresponding to the painting element in the restricted element group can be extracted from the retrospective frame, and the contour can be determined as a lossless contour. The lossless contour refers to a complete and clear painting contour corresponding to the painting element in the restricted element group in the retrospective frame.

[0087] B43, calling preset lossless pixel values ​​to update the image area except the lossless contour in the traceback frame to obtain the lossless frame corresponding to the restricted element group, and replacing and updating the shape display frame in the restricted element group according to the lossless frame.

[0088] In practical applications, the retrospective frame contains complete and clear painting contours corresponding to all graffiti element groups, that is, the retrospective frame contains the painting contours corresponding to the damaged element group. Since the painting contours in the damaged element group may be blocked by the hand contour, after determining the lossless contours corresponding to the painting contours in the damaged element group in the retrospective frame, the preset lossless pixel values ​​can be called to update the image area except the lossless contours in the retrospective frame, so as to retain the clear lossless contours of the corresponding painting elements. After obtaining the lossless frame corresponding to the restricted element group, the shape display frame that does not meet the display conditions in the restricted element group can be replaced with the new lossless frame, so as to ensure that the painting elements can be displayed completely and clearly.

[0089] The preset lossless pixel value refers to a preset pixel value corresponding to the background color of the canvas. For example, when the background color of the canvas is white, the preset lossless pixel value may be a pixel value corresponding to white.

[0090] Through the above implementation, it can be ensured that the image data can be more complete and clearer, thereby improving the accuracy of the acquired graffiti data.

[0091] S3, constructing a graffiti timeline corresponding to the graffiti recording video, and generating a plurality of graffiti recording nodes corresponding to each of the graffiti element groups on the graffiti timeline to obtain a graffiti recording axis.

[0092] It is understandable that when displaying the graffiti data of the painter during the graffiti process, this solution is displayed through the graffiti record axis. Specifically, a graffiti timeline corresponding to the graffiti recording video can be constructed, and the timeline represents the time process from the beginning to the end of the graffiti recording video. On the graffiti timeline, each graffiti element group can construct a corresponding graffiti recording node at its corresponding graffiti moment to form a graffiti recording axis, which can record its complete graffiti process. Among them, the graffiti timeline refers to the timeline of the recording time period corresponding to the graffiti recording video, the graffiti recording node refers to the node that can display the graffiti feature data, and the graffiti recording axis refers to the axis that records the graffiti feature data at different times.

[0093] In some embodiments, the scribble record axis may be generated in the following manner: S31, constructing a corresponding graffiti timeline according to the recording time corresponding to the graffiti recording video, obtaining the recording moment corresponding to each key frame in each graffiti element group, and determining the graffiti time point corresponding to each recording moment in the graffiti timeline.

[0094] In practical applications, first, the corresponding graffiti timeline can be constructed according to the recording time period corresponding to the graffiti recording video. The starting point of the graffiti timeline corresponds to the start time of the graffiti recording video, and the end point corresponds to the end time of the graffiti recording video. For example, if the painter graffiti in the time period from 9 to 10 in the morning, then a graffiti timeline from 9 to 10 can be constructed. Since the recording time corresponding to the key frame can be obtained when the key frame is obtained, the position point corresponding to the recording time can be found on the constructed graffiti timeline, that is, the graffiti time point. Among them, the graffiti time point refers to the time point when the graffiti feature is completely drawn on the graffiti timeline.

[0095] S32, constructing an initial recording node of a shape display type at the graffiti time point corresponding to the shape display frame, and constructing an initial recording node of a coloring display type at the graffiti time point corresponding to the coloring display frame.

[0096] Specifically, at the graffiti time point corresponding to the shape display frame on the graffiti timeline, an initial recording node can be constructed, and the node can be marked as a shape display type, which can indicate that the node corresponds to the graffiti shape feature. At the graffiti time point corresponding to the coloring display frame on the graffiti timeline, an initial recording node can also be constructed, and the node can be marked as a coloring display type, which can indicate that the node corresponds to the graffiti color feature. Among them, the initial recording node refers to the initial node corresponding to each key frame.

[0097] S33, updating the shape of the corresponding initial record node according to the shape of the drawing outline in the shape display frame to obtain a graffiti record node corresponding to the shape display frame.

[0098] Specifically, the outline of the painting element can be extracted from the shape display frame, and the shape of the initial recording node of the shape display type can be updated according to the shape of the painting outline, so as to obtain the graffiti recording node corresponding to the shape display frame, and then the graffiti shape at the corresponding time point can be recorded.

[0099] See also Figure 3 , is a schematic diagram of a graffiti recording axis provided by an embodiment of the present invention, such as Figure 3 As shown in , the graffiti record axis includes 4 graffiti record nodes. The first graffiti record node and the third graffiti record node are graffiti record nodes corresponding to the shape display frame. The shapes of the first and third graffiti record nodes can be updated by the shape of the drawing outline. It can be seen that the graffiti shape of the shape display frame corresponding to the first graffiti record node is the sun, and the graffiti shape of the shape display frame corresponding to the third graffiti record node is the star.

[0100] S34, updating the color of the corresponding initial record node based on the filling color corresponding to the color-filled display frame to obtain a graffiti record node corresponding to the color-filled display frame, and updating the graffiti timeline according to the graffiti record node to obtain a graffiti record axis.

[0101] Specifically, the corresponding filling color can be obtained from the coloring display frame, and the color of the initial recording node of the coloring display type is updated according to the filling color, and the graffiti recording node corresponding to the coloring display type can be obtained, so that the coloring data at the corresponding moment can be recorded.

[0102] For example Figure 3 As shown in , the second graffiti record node and the fourth graffiti record node on the graffiti record axis are graffiti record nodes corresponding to the coloring display frame. The filling color of the coloring display frame corresponding to the second graffiti record node is yellow, and the filling color of the coloring display frame corresponding to the fourth graffiti record node is red. By updating the graffiti record nodes in combination with the image data, the following can be obtained: Figure 3The graffiti recording axis shown is capable of recording complete graffiti data.

[0103] Through the above implementation, the graffiti data recorded at different times during the drawing process can be combined on the graffiti record axis to improve the comprehensiveness of data analysis.

[0104] S4, inputting the graffiti record axis into a graffiti analysis model to obtain graffiti analysis data.

[0105] The graffiti analysis model is a model used to analyze image data in the graffiti process, and the graffiti analysis data refers to data obtained after analyzing the graffiti record axis.

[0106] Based on the above embodiment, the specific implementation of step S4 may be: S41, retrieving preset graffiti features based on the graffiti analysis model, performing feature comparison on a plurality of graffiti record nodes in the graffiti record axis according to the preset graffiti features, and determining graffiti features corresponding to each of the graffiti record nodes.

[0107] The preset graffiti features include preset shape features and preset color features, and each of the preset graffiti features includes a corresponding preset graffiti icon.

[0108] In actual applications, in the graffiti analysis model, a series of graffiti features can be preset, such as pentagon, yellow, etc., and different graffiti features correspond to different graffiti icons. For example, the graffiti icon corresponding to the five-pointed star can be a smiling face, and the graffiti icon corresponding to the yellow can be a crying face. The graffiti icons can show the emotional changes of the painter at different times, providing managers with more comprehensive data references.

[0109] Specifically, feature comparison can be performed on each graffiti record node in the graffiti record axis according to preset graffiti features to determine which preset feature the shape feature or color feature of each graffiti record node matches, and then the graffiti feature corresponding to the graffiti record node can be determined. The graffiti icon corresponding to the node can be determined based on the graffiti feature, and subsequently the emotional reference data at different times can be intuitively displayed on the graffiti record axis through the graffiti icon.

[0110] The preset graffiti features refer to preset features that can be used to identify graffiti shapes, and the preset graffiti icons refer to icons that can be used to intuitively display graffiti features.

[0111] S42, updating the graffiti record axis according to the preset graffiti icons corresponding to the graffiti features to obtain graffiti analysis data corresponding to the target identity.

[0112] Specifically, after obtaining the graffiti feature corresponding to each graffiti record node, the graffiti record axis can be updated through the preset graffiti icon corresponding to the graffiti feature, so that the emotion changes at different moments can be displayed through the icon.

[0113] In some embodiments, multiple graffiti icons corresponding to a graffiti element group may have certain correlations. Therefore, when multiple graffiti icons corresponding to a graffiti element group are correlated, the correlated graffiti icons can be framed by the same display frame, and when multiple correlated graffiti icons correspond to joint icons, the joint image can be displayed at a preset position of the display frame.

[0114] For example, when a smiling face icon and a surprised face icon may correspond to a surprise face icon, in this case, the smiling face icon and the surprised face icon may be framed, and the combined information of the two icons may be displayed through the surprise face icon.

[0115] Through the above implementation, the specific graffiti icon corresponding to each graffiti record node can be accurately identified, so that the graffiti data of the painter in the corresponding time period can be accurately obtained, thereby improving the comprehensiveness of data display.

[0116] Figure 4 The apparatus of the embodiment shown can be used to perform Figure 1 The implementation principles and technical effects of the steps in the method embodiment shown are similar and will not be repeated here.

[0117] See also Figure 4 , is a structural diagram of an AI-driven graffiti painting emotional psychological state analysis and recognition data processing system provided by an embodiment of the present invention. The data processing system based on the AI-driven graffiti painting emotional psychological state analysis and recognition data processing system includes: A calling module, used to call a graffiti recording video corresponding to a target identity input by a management terminal, and determine a plurality of key frames corresponding to the graffiti recording video according to a feature analysis model; An acquisition module, used for acquiring the painting elements of each key frame, and dividing each key frame according to the painting elements to obtain a plurality of graffiti element groups; A construction module, used to construct a graffiti timeline corresponding to the graffiti recording video, and generate a plurality of graffiti record nodes corresponding to each of the graffiti element groups on the graffiti timeline to obtain a graffiti record axis; The analysis module is used to input the graffiti record axis into a graffiti analysis model to obtain graffiti analysis data.

[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An AI-driven graffiti painting emotional psychological state analysis and recognition data processing method, characterized in that: include: Retrieving a graffiti recording video corresponding to a target identity input by a management terminal, and determining a plurality of key frames corresponding to the graffiti recording video according to a feature analysis model; Acquire the painting elements of each of the key frames, and divide each of the key frames according to the painting elements to obtain a plurality of graffiti element groups; Constructing a graffiti timeline corresponding to the graffiti recording video, and generating a plurality of graffiti recording nodes corresponding to each of the graffiti element groups on the graffiti timeline to obtain a graffiti recording axis; The graffiti record axis is input into a graffiti analysis model to obtain graffiti analysis data.

2. The method according to claim 1, characterized in that Retrieving a graffiti recording video corresponding to a target identity input by the management end, and determining a plurality of key frames corresponding to the graffiti recording video according to a feature analysis model, including: Arranging the recorded frames in the graffiti recorded video in chronological order to obtain a screening sequence, and sequentially acquiring the painting contours in the adjacent recorded frames in the screening sequence based on the feature analysis model; Acquire multiple drawing points corresponding to the drawing outline, perform coordinate processing on the recording frame, and obtain drawing coordinates of each drawing point; Graffiti feature recognition is performed on the recording frame according to the drawing coordinates and the drawing pixel values ​​corresponding to each drawing point, and a recording frame satisfying a graffiti feature condition is determined as a key frame, wherein the graffiti feature condition includes a shape feature condition and a coloring feature condition.

3. The method according to claim 2, characterized in that According to the drawing coordinates and drawing pixel values ​​corresponding to each drawing point, the recording frame is identified with graffiti features, and a recording frame satisfying a graffiti feature condition is determined as a key frame, wherein the graffiti feature condition includes a shape feature condition and a coloring feature condition, including: Determine that a drawing point in each recording frame whose drawing coordinates are consistent with those in a previous recording frame is an existing drawing point, and a drawing point in each recording frame whose drawing coordinates are inconsistent with those in a previous recording frame is a newly added drawing point; When the newly added drawing point does not exist in the recording frames of the preset number of consecutive frames, determining that the last recording frame of the consecutive multiple recording frames meets the shape feature condition, and obtaining the corresponding recording frame as a key frame; Obtaining the interval distance between the closest existing drawing point and the newly added drawing point in each of the recording frames; When the interval distance is greater than or equal to the shape filling threshold and / or the drawing pixel value of the newly added drawing point is inconsistent with the existing drawing point, it is determined that the previous recording frame adjacent to the recording frame meets the color filling feature condition, the corresponding recording frame is obtained as a key frame, and the drawing pixel value corresponding to the existing drawing point in the key frame is determined to be the filling color corresponding to the key frame.

4. The method according to claim 3, characterized in that Also includes: Arranging a plurality of key frames satisfying the shape feature conditions according to a time sequence to obtain a shape screening sequence; Sequentially acquiring adjacent key frames in the shape screening sequence as target frames, and extracting target contours in the target frames; Determine that the drawing points in each target outline where the drawing coordinates and the drawing pixel values ​​are consistent are the same drawing points, and count the same number of the same drawing points in the previous target frame; The matching similarity is obtained according to the ratio between the same number and the total number of all drawing points in the previous target frame. When the matching similarity is greater than a similarity threshold, the previous target frame is deleted.

5. The method according to claim 3, characterized in that: Acquiring the painting elements of each of the key frames, and dividing each of the key frames according to the painting elements to obtain a plurality of graffiti element groups, including: Performing element feature recognition on the painting outline in each of the key frames according to preset element features, and determining the painting element corresponding to each of the key frames; The key frames with the same painting elements are obtained as the same graffiti element group, and the key frames in the graffiti element group that meet the shape feature condition are determined as shape display frames, and the key frames that meet the coloring feature condition are determined as coloring display frames.

6. The method according to claim 5, characterized in that Also includes: Performing hand contour recognition on each of the recorded frames, and determining the recorded frames without the hand contour as pending frames; Arranging the pending frames in reverse time order to obtain a backtracking sequence, and sequentially identifying elements of the pending frames in the backtracking sequence; When the pending frame has drawing elements corresponding to all graffiti element groups, and the clarity of the pending frame is greater than or equal to a clarity threshold, determining that the pending frame is a traceback frame; A graffiti element group whose shape display frame does not meet the shape display condition is obtained as a restricted element group, a lossless frame corresponding to the restricted element group is generated according to the backtracking frame, and the restricted element group is updated based on the lossless frame.

7. The method according to claim 6, characterized in that Acquiring a graffiti element group whose shape display frame does not meet the shape display condition as a restricted element group, generating a lossless frame corresponding to the restricted element group according to the backtracking frame, and updating the restricted element group based on the lossless frame, including: When the shape display frame has the hand outline and / or the clarity of the shape display frame is less than the clarity threshold, determining that the shape display frame does not meet the shape display condition, and determining that the graffiti element group corresponding to the shape display frame is a restricted element group; Acquire a drawing contour corresponding to the drawing element of the restricted element group in the traceback frame as a lossless contour; The preset lossless pixel value is retrieved to update the image area except the lossless contour in the traceback frame to obtain the lossless frame corresponding to the restricted element group, and the shape display frame in the restricted element group is replaced and updated according to the lossless frame.

8. The method according to claim 5, characterized in that Constructing a graffiti timeline corresponding to the graffiti recording video, and generating a plurality of graffiti recording nodes corresponding to each of the graffiti element groups on the graffiti timeline to obtain a graffiti recording axis, including: Constructing a corresponding graffiti timeline according to the recording time corresponding to the graffiti recording video, obtaining the recording time corresponding to each key frame in each graffiti element group, and determining the graffiti time point corresponding to each recording time in the graffiti timeline; Constructing an initial recording node of a shape display type at a graffiti time point corresponding to the shape display frame, and constructing an initial recording node of a coloring display type at a graffiti time point corresponding to the coloring display frame; Performing shape update on the corresponding initial record node according to the shape of the drawing outline in the shape display frame to obtain a graffiti record node corresponding to the shape display frame; The color of the corresponding initial record node is updated based on the filling color corresponding to the coloring display frame to obtain the graffiti record node corresponding to the coloring display frame, and the graffiti timeline is updated according to the graffiti record node to obtain the graffiti record axis.

9. The method according to claim 8, characterized in that Inputting the graffiti record axis into a graffiti analysis model to obtain graffiti analysis data includes: Retrieving preset graffiti features based on the graffiti analysis model, performing feature comparison on a plurality of graffiti record nodes in the graffiti record axis according to the preset graffiti features, and determining graffiti features corresponding to each of the graffiti record nodes; Wherein, the preset graffiti features include preset shape features and preset color features, and each of the preset graffiti features includes a corresponding preset graffiti icon; The graffiti record axis is updated according to the preset graffiti icons corresponding to the graffiti features to obtain graffiti analysis data corresponding to the target identity.

10. An AI-driven graffiti painting emotional psychological state analysis and recognition data processing system, characterized in that: include: A calling module, used to call a graffiti recording video corresponding to a target identity input by a management terminal, and determine a plurality of key frames corresponding to the graffiti recording video according to a feature analysis model; An acquisition module, used for acquiring the painting elements of each key frame, and dividing each key frame according to the painting elements to obtain a plurality of graffiti element groups; A construction module, used to construct a graffiti timeline corresponding to the graffiti recording video, and generate a plurality of graffiti record nodes corresponding to each of the graffiti element groups on the graffiti timeline to obtain a graffiti record axis; The analysis module is used to input the graffiti record axis into a graffiti analysis model to obtain graffiti analysis data.

Images