Art design evaluation system based on data analysis

Abstract

The invention discloses an artistic design evaluation system based on data analysis, and relates to the technical field of artistic evaluation, and the system comprises the steps: collecting work data, carrying out image feature extraction, outputting scoring reference data, inputting the scoring reference data to a work scoring model, outputting the scoring data, segmenting the current work image data according to the scoring data, and carrying out the segmentation of the current work image data. According to the method, objectivity and professionality of artistic design evaluation are improved through data analysis, multi-dimensional features such as image features, sentiment analysis and historical data are combined to form a scientific scoring model, and meanwhile, improvement suggestions are provided for works with relatively low scores. Reference is provided for artist optimization design in combination with similar features of excellent cases, evaluation is more comprehensive and profound through multi-dimensional feature analysis, improvement recommendation supports work quality improvement, and the method is suitable for objective evaluation of large-scale works of auction offices, artworks and the like and has great application value.

Description

Technical Field

[0001] The present invention relates to the technical field of art evaluation, and in particular to an art evaluation system based on data analysis. Background Art

[0002] With the advancement of artificial intelligence and big data analysis technology, data-driven evaluation systems have gradually been introduced into the field of art design. Traditional art design evaluation usually relies on the subjective judgment of experts. Although this method has its unique value, it is easily affected by subjective biases and cognitive limitations, which makes it difficult to ensure the accuracy and objectivity of the evaluation results. Therefore, art design evaluation systems based on data analysis have gradually become an important exploration direction to improve the objectivity and depth of evaluation of artworks.

[0003] At present, the Chinese invention patent with application number CN202410627761.3 discloses an art design evaluation system and method based on big data. By collecting, organizing and analyzing art design related data, it helps evaluators to evaluate and analyze works. This system can establish an evaluation model for art design works through machine learning and data mining and other technologies. By analyzing and learning a large amount of sample data, the system can learn the laws and characteristics of the field of art design, so that it can evaluate and predict unknown works. Using this system, users can input art design works to be evaluated, and the system can analyze and evaluate them according to existing models, and give evaluations on the characteristics, style, innovation, etc. of the works. At the same time, the system can also provide information such as other works, artists, and historical background related to the work to help users have a more comprehensive understanding and evaluation.

[0004] The above-mentioned technology makes it difficult to provide references for improving art designs while scoring them, which affects the professionalism and richness of the evaluation results of artworks. Summary of the invention

[0005] The technical problem solved by the present invention is that the above-mentioned technology is difficult to provide references for improving artistic designs while scoring them, which affects the professionalism and richness of the evaluation results of the artworks.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: an art design evaluation system based on data analysis, comprising a data collection and processing module, a feature extraction and analysis module, a scoring model design module and a case selection and recommendation module;

[0007] The data collection and processing module is used to collect work data, which includes current work image data, current work text evaluation data, current work rating data set and current artist historical auction price data set, and pre-process the work data;

[0008] The feature extraction and analysis module is used to extract image features from the current work image data, output the current image feature data, input the current work text evaluation data into the sentiment analysis model, output the current sentiment analysis data, calculate the current work score mean according to the current work score data set, output the current work score data, calculate the current artist historical auction price mean according to the current artist historical auction price data set, output the current artist historical auction price mean data, and combine the current image feature data, the current sentiment analysis data, the current work score data and the current artist historical auction price mean data to output as the score reference data;

[0009] The scoring model design module is used to establish a work scoring model, input scoring reference data into the work scoring model, and output scoring data;

[0010] The case selection recommendation module is used to segment the image data of the current work according to the scoring data, output it as a number of segmented image data, perform feature extraction on the segmented image data, output it as segmented image feature data, establish a first correspondence between the segmented image data and the segmented image feature data, input the segmented image feature data into a preset database for matching, output the matching results, establish a second correspondence between the segmented image data and the matching results, and output it as image improvement suggestion data.

[0011] Preferably, the data collection and processing module includes a data collection unit and a data processing unit;

[0012] The data collection unit is used to collect the current work image data, the current work text evaluation data, the current work score data set and the current artist historical auction price data set, and merge the current work image data, the current work text evaluation data, the current work score data set and the current artist historical auction price data set to output as work data;

[0013] The data processing unit is used to pre-process the work data, and the pre-processing is to perform word segmentation processing, stop word removal processing and duplicate removal processing on the current work text evaluation data.

[0014] Preferably, the feature extraction and analysis module includes a data analysis unit, a data calculation unit and a data integration unit;

[0015] The data analysis unit is used to extract image features from the current work image data, and the image feature extraction process is as follows:

[0016] Convert the current work image data to Lab color space, calculate the color histogram of the current work image data in Lab color space, obtain the pixel ratio of each color, output as a color ratio array, calculate the mean, and the mathematical expression is:

[0017]

[0018] Among them, μ is the mean, p k is the kth color ratio, k is an integer greater than 1, and n is the total number of non-zero color intervals;

[0019] Calculate the standard deviation, the mathematical expression is:

[0020]

[0021] Where σ is the standard deviation;

[0022] Calculate the maximum standard deviation, the mathematical expression is:

[0023]

[0024] Among them, σ max is the maximum standard deviation;

[0025] Calculate the richness, the mathematical expression is:

[0026]

[0027] Use the edge detection algorithm to extract the edge lines of the current work image data, obtain the pixel position of the edge lines, calculate the gradient angle data, divide the gradient angle data into 18 intervals with 10 degrees as the gradient value, construct the direction histogram, and calculate the information entropy of the direction distribution. The mathematical expression is:

[0028] Entropy=-∑p i log(p i )

[0029] Among them, Entropy is information entropy, p i is the probability distribution of the i-th interval in the direction histogram;

[0030] Combine the richness and information entropy inputs and output them as the current image feature data;

[0031] The current work text evaluation data is input into the sentiment analysis model. The logic of the sentiment analysis model is:

[0032] Use SentiWordNet to perform sentiment annotation on the text evaluation data of the current work. The SentiWordNet includes positive words, scores corresponding to positive words, negative words, scores corresponding to negative words, neutral words and scores corresponding to neutral words. Obtain the scores corresponding to the positive words and the scores corresponding to the negative words of the text evaluation data of the current work, and calculate the positive weight and the negative weight respectively. The mathematical expression is:

[0033]

[0034] The sentiment tendency value is calculated based on the positive weight and negative weight. The mathematical expression is:

[0035] Sentiment tendency value = positive weight + negative weight

[0036] Output the sentiment tendency value as the current sentiment analysis data.

[0037] Preferably, the data calculation unit is used to calculate the current work score mean of the current work score data set, and output it as the current work score mean data, and calculate the current artist historical auction price mean of the current artist historical auction price data set, and output it as the current artist historical auction price mean data;

[0038] The data integration unit is used to combine the current image feature data, the current sentiment analysis data, the current work score average data and the current artist historical auction price average data to output as score reference data.

[0039] Preferably, the scoring model design module includes a model building unit and a model output unit;

[0040] The model building unit is used to build a work analysis model, and the logic of the work analysis model is:

[0041] Receive image feature data, sentiment analysis data, work rating average data, and historical auction price average data;

[0042] Weight the richness, information entropy, sentiment value, work rating average data, and historical auction price average data:

[0043] The weight of sentiment value is 0.25, the weight of richness is 0.2, the weight of mean data of work rating is 0.2, the weight of mean data of historical auction price is 0.2, and the weight of information entropy is 0.15;

[0044] Calculate the total weight of the richness, information entropy, sentiment tendency value, work rating mean data and historical auction price mean data, and output it as rating data;

[0045] The model output unit is used to input the scoring reference data into the work scoring model, obtain the scoring data, and output the current work scoring data.

[0046] Preferably, the case selection recommendation module includes an image segmentation unit, an image matching unit and an improved output unit;

[0047] The image segmentation unit is used to obtain the current work rating data. If the current work rating data is lower than the preset average work rating data, the Mask RCNN model is used to segment the current work image data and output it as a number of segmented image data. Image feature extraction is performed on the segmented image data. The image feature extraction process is the same as that of the data analysis unit. The richness and information entropy are obtained and output as segmented image richness and segmented image information entropy. The segmented image richness and segmented image information entropy are combined and output as segmented image feature data, and a first corresponding relationship between the segmented image data and the segmented image feature data is established.

[0048] Preferably, the image matching unit is used to input the segmented image feature data into a database, wherein the database includes case segmented image data, case scoring scores corresponding to the case segmented images, case segmented image richness and case segmented image information entropy, calculate the cosine similarity between the segmented image richness and the case segmented image richness, and output it as richness similarity data, calculate the cosine similarity between the segmented image information entropy and the case segmented image information entropy, and output it as complexity similarity data, calculate the sum of the richness similarity data and the complexity similarity data, and output it as total similarity data, select case segmented image data whose total similarity data is greater than a preset total similarity threshold, and output it as a similar image data set as a matching result.

[0049] Preferably, the improvement output unit is used to select the case segmentation image data with the highest case score in the matching results, establish a second corresponding relationship between the segmented image data and the matching results, and output it as image improvement suggestion data.

[0050] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the art design evaluation system based on data analysis is implemented.

[0051] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the art design evaluation system based on data analysis is implemented.

[0052] Beneficial effects of the invention: The invention improves the objectivity and professionalism of art design evaluation through data analysis, combines multi-dimensional features such as image features, sentiment analysis, historical data, etc. to form a scientific scoring model. At the same time, it provides improvement suggestions for works with low scores, and provides a reference for artists to optimize their designs based on similar features of excellent cases. Multi-dimensional feature analysis makes the evaluation more comprehensive and profound, and improved recommendations support the improvement of work quality. It is suitable for the objective evaluation of large-scale works in auction houses, art galleries, etc., and has great application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A basic flow chart of an art design evaluation system based on data analysis is provided as an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0055] Example, see Figure 1 , provides an art design evaluation system based on data analysis, including a data collection and processing module, a feature extraction and analysis module, a scoring model design module and a case selection and recommendation module.

[0056] The data collection and processing module is used to collect work data, which includes current work image data, current work text evaluation data, current work rating data set and current artist historical auction price data set, and pre-process the work data.

[0057] The feature extraction and analysis module is used to extract image features from the current work image data, output the current image feature data, input the current work text evaluation data into the sentiment analysis model, output the current sentiment analysis data, calculate the current work score average based on the current work score data set, output the current work score data, calculate the current artist's historical auction price average based on the current artist's historical auction price data set, output the current artist's historical auction price average data, and combine the current image feature data, current sentiment analysis data, current work score data and current artist's historical auction price average data to output as scoring reference data.

[0058] The scoring model design module is used to establish a work scoring model, input scoring reference data into the work scoring model, and output scoring data.

[0059] The case selection recommendation module is used to segment the image data of the current work according to the scoring data, output it as a number of segmented image data, perform feature extraction on the segmented image data, output it as segmented image feature data, establish a first correspondence between the segmented image data and the segmented image feature data, input the segmented image feature data into a preset database for matching, output the matching results, establish a second correspondence between the segmented image data and the matching results, and output it as image improvement suggestion data.

[0060] The data collection and processing module includes a data collection unit and a data processing unit.

[0061] The data collection unit is used to collect the current work image data, the current work text evaluation data, the current work rating data set and the current artist historical auction price data set, and merge the current work image data, the current work text evaluation data, the current work rating data set and the current artist historical auction price data set to output as work data.

[0062] The data collection unit ensures the integrity of the work data by collecting various information such as image data, text evaluation data, rating data, and historical auction data. It integrates data from different sources into a unified work data output, provides comprehensive data support for subsequent analysis, and ensures that the evaluation content has multi-dimensional reference.

[0063] The data processing unit is used to pre-process the work data, and the pre-processing is to perform word segmentation processing, stop word removal processing and duplicate removal processing on the current work text evaluation data.

[0064] The data processing unit preprocesses the text evaluation data in the work data, including word segmentation, stop word removal and deduplication, to ensure the conciseness and accuracy of the text data. The preprocessing step removes noise words, improves the clarity and usability of the data, and thus improves the accuracy of sentiment analysis and feature extraction.

[0065] The data collection and processing module realizes the comprehensive collection and preprocessing of the work data, ensuring that the input data information is complete and clean, providing a high-quality data foundation for subsequent analysis and scoring. Through the integration of image data, text evaluation data, scoring data and artist historical data, it comprehensively reflects the multi-dimensional information of the work and its evaluation, laying a data foundation for accurate evaluation.

[0066] The feature extraction and analysis module includes a data analysis unit, a data calculation unit and a data integration unit.

[0067] The data analysis unit is used to extract image features from the current work image data. The image feature extraction process is as follows:

[0068] Convert the current work image data to Lab color space, calculate the color histogram of the current work image data in Lab color space, obtain the pixel ratio of each color, output as a color ratio array, calculate the mean, and the mathematical expression is:

[0069]

[0070] Among them, μ is the mean, p k is the kth color ratio, k is an integer greater than 1, and n is the total number of non-zero color intervals.

[0071] Calculate the standard deviation, the mathematical expression is:

[0072]

[0073] Here, σ is the standard deviation.

[0074] Calculate the maximum standard deviation, the mathematical expression is:

[0075]

[0076] Among them, σ max is the maximum standard deviation.

[0077] Calculate the richness, the mathematical expression is:

[0078]

[0079] Use the edge detection algorithm to extract the edge lines of the current work image data, obtain the pixel position of the edge lines, calculate the gradient angle data, divide the gradient angle data into 18 intervals with 10 degrees as the gradient value, construct the direction histogram, and calculate the information entropy of the direction distribution. The mathematical expression is:

[0080] Entropy=-∑p i log(p i )

[0081] Among them, Entropy is information entropy, p i is the probability distribution of the i-th interval in the direction histogram.

[0082] The richness and information entropy are combined and output as the current image feature data.

[0083] The current work text evaluation data is input into the sentiment analysis model. The logic of the sentiment analysis model is:

[0084] Use SentiWordNet to perform sentiment annotation on the text evaluation data of the current work. The SentiWordNet includes positive words, scores corresponding to positive words, negative words, scores corresponding to negative words, neutral words and scores corresponding to neutral words. Obtain the scores corresponding to the positive words and the scores corresponding to the negative words of the text evaluation data of the current work, and calculate the positive weight and the negative weight respectively. The mathematical expression is:

[0085]

[0086] The sentiment tendency value is calculated based on the positive weight and negative weight. The mathematical expression is:

[0087] Sentiment tendency value = positive weight + negative weight

[0088] Output the sentiment tendency value as the current sentiment analysis data.

[0089] The data analysis unit is responsible for extracting image features and sentiment data, calculating the image's color ratio, richness, information entropy and other feature data, and capturing the image's structural information through edge detection. At the same time, it performs sentiment analysis on text evaluation data, calculates positive and negative sentiment tendencies, and outputs sentiment data, providing detailed feature and sentiment analysis results for images and texts, and providing multi-dimensional support for scoring.

[0090] The data calculation unit is used to calculate the current work score mean of the current work score data set, and output it as the current work score mean data, and calculate the current artist historical auction price mean of the current artist historical auction price data set, and output it as the current artist historical auction price mean data.

[0091] The data calculation unit calculates the mean score of the work rating data set and the mean of the artist's historical auction prices, and outputs the mean score data and the mean historical price data. The statistical information provided reflects the recognition and value of the works and artists in the market, and provides a data-level reference for the rating.

[0092] The data integration unit is used to combine the current image feature data, the current sentiment analysis data, the current work score average data and the current artist historical auction price average data and output them as score reference data.

[0093] The data integration unit combines image features, sentiment analysis, average ratings and average historical auction prices into rating reference data, ensuring that the input data of the rating model is multi-dimensional, rich and accurate, providing complete and objective data support for the final rating.

[0094] The feature extraction and analysis module extracts image features, sentiment data and statistical data through in-depth analysis of various data such as work images, text reviews, ratings and historical prices, ensuring the comprehensiveness and accuracy of multi-dimensional information in the evaluation process. By integrating image features and sentiment analysis results as scoring reference data, the module provides complete data support for the scoring model, making the evaluation of artworks more scientific, objective and accurate.

[0095] The scoring model design module includes a model building unit and a model output unit.

[0096] The model building unit is used to build a work analysis model, the logic of which is:

[0097] Receive image feature data, sentiment analysis data, work rating average data, and historical auction price average data.

[0098] Weight the richness, information entropy, sentiment value, work rating average data, and historical auction price average data:

[0099] The weight value of sentiment tendency value is 0.25, the weight value of richness is 0.2, the weight value of mean data of work rating is 0.2, the weight value of mean data of historical auction price is 0.2, and the weight value of information entropy is 0.15.

[0100] Calculate the total weight of richness, information entropy, sentiment tendency value, work rating mean data and historical auction price mean data, and output it as rating data.

[0101] The model building unit is responsible for constructing a scoring model and performing weighted processing based on image features, sentiment analysis, scoring averages and historical auction data to ensure that the weight of each feature's impact on the score is reasonably distributed. The design ensures that the scoring model is scientific and can more accurately reflect the multi-faceted value of the work.

[0102] The model output unit is used to input the scoring reference data into the work scoring model, obtain the scoring data, and output the current work scoring data.

[0103] The model output unit inputs the scoring reference data into the scoring model to generate the final scoring data of the current work. By processing and outputting the scoring reference data, the final numerical expression is provided for the scoring result, making the evaluation result of the work concise, clear, easy to understand and reference.

[0104] The scoring model design module realizes the quantitative scoring of works. By establishing a scoring model that includes multiple features, it integrates image features, emotional tendencies, scoring means and historical price data in a weighted manner to make the scoring results more objective and professional. It outputs a comprehensive score based on the weights of different features, providing a scientific and standard scoring basis for the market and evaluation system of artworks.

[0105] The case selection recommendation module includes an image segmentation unit, an image matching unit and an improved output unit.

[0106] The image segmentation unit is used to obtain the current work rating data. If the current work rating data is lower than the preset average work rating data, the MaskRCNN model is used to segment the current work image data and output it as a number of segmented image data. Image feature extraction is performed on the segmented image data. The image feature extraction process is the same as that of the data analysis unit. The richness and information entropy are obtained and output as segmented image richness and segmented image information entropy. The segmented image richness and segmented image information entropy are combined and output as segmented image feature data, and a first corresponding relationship between the segmented image data and the segmented image feature data is established.

[0107] When the score of a work is lower than the preset standard, the image segmentation unit divides the work image into several regions, extracts features such as the richness and information entropy of each part, and forms segmented image feature data. Through refinement processing, the image segmentation unit provides fine feature support for the improvement of low-scoring works, making subsequent analysis more accurate.

[0108] The image matching unit is used to input the segmented image feature data into a database, wherein the database includes case segmented image data, case scoring scores corresponding to the case segmented images, case segmented image richness and case segmented image information entropy, calculate the cosine similarity between the segmented image richness and the case segmented image richness, and output it as richness similarity data, calculate the cosine similarity between the segmented image information entropy and the case segmented image information entropy, and output it as complexity similarity data, calculate the sum of the richness similarity data and the complexity similarity data, and output it as total similarity data, select the case segmented image data whose total similarity data is greater than a preset total similarity threshold, and output it as a similar image data set as a matching result.

[0109] The image matching unit inputs the segmented image feature data into the database, calculates the cosine similarity with excellent cases, and outputs case data with higher similarity. By calculating the similarity data of richness and information entropy, the image matching unit accurately selects high-quality cases with high similarity to the current work, providing strong reference data for improvement suggestions.

[0110] The improved output unit is used to select the case segmentation image data with the highest case score in the matching results, establish a second corresponding relationship between the segmented image data and the matching results, and output it as image improvement suggestion data.

[0111] The improved output unit selects the highest-scoring case segmentation image data from the matching results to form specific improvement suggestions for the current work. Through the recommendation, artists can refer to excellent cases similar to their own works, thereby obtaining practical optimization directions and improving the overall performance of their works.

[0112] The case selection recommendation module provides optimization suggestions for low-scoring works through multi-step analysis of image segmentation, matching, and improvement suggestions. The module segments images based on MaskRCNN, extracts features from each segmented part, and compares similarities with excellent cases in the database to generate guiding improvement suggestions, effectively supporting design optimization and improvement, and providing practical and data-driven improvement references.

[0113] The present invention introduces a variety of data analysis methods, combines the image features of the works with sentiment analysis, and obtains scoring data through comprehensive calculation, which reduces subjective evaluation bias and improves the objectivity of the evaluation. The scoring model makes a fine setting of the scoring weights of the works, so that multi-dimensional features such as richness and information entropy are reasonably distributed in the final score, thereby improving the professionalism and scientificity of the evaluation results. When the score is low, the case selection recommendation module will perform fine-grained segmentation on the works, extract the feature data of each part, and search for excellent cases with high similarity in the database. By comparing the matching results with the cases, the system can put forward practical improvement suggestions to help artists optimize the design in subsequent creations. This process effectively supports the improvement of the quality of the works, so that design improvements no longer rely on pure subjective evaluation, and integrates multiple feature indicators to make the evaluation results more three-dimensional and comprehensive. At the same time, the text sentiment analysis model further enriches the evaluation system by calculating positive and negative weights, revealing the potential emotional impact of the work on the viewer's mind. Through multi-dimensional and multi-level feature analysis, the system effectively enhances the depth of interpretation of art design. The adopted scoring model is not only suitable for the evaluation of a single work, but also has large-scale, automated evaluation capabilities, and is suitable for batch evaluation of works in auction houses, art galleries and other scenarios.

[0114] It should be understood by those skilled in the art that the embodiments of the present invention can be provided as methods, systems or computer program products. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program codes. Among them, the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic memory, flash memory, disk or optical disk. These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0115] 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 preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An art design evaluation system based on data analysis, characterized in that: It includes data collection and processing module, feature extraction and analysis module, scoring model design module and case selection and recommendation module; The data collection and processing module is used to collect work data, which includes current work image data, current work text evaluation data, current work rating data set and current artist historical auction price data set, and pre-process the work data; The feature extraction and analysis module is used to extract image features from the current work image data, output the current image feature data, input the current work text evaluation data into the sentiment analysis model, output the current sentiment analysis data, calculate the current work score mean according to the current work score data set, output the current work score data, calculate the current artist historical auction price mean according to the current artist historical auction price data set, output the current artist historical auction price mean data, and combine the current image feature data, the current sentiment analysis data, the current work score data and the current artist historical auction price mean data to output as the score reference data; The scoring model design module is used to establish a work scoring model, input scoring reference data into the work scoring model, and output scoring data; The case selection recommendation module is used to segment the image data of the current work according to the scoring data, output it as a number of segmented image data, perform feature extraction on the segmented image data, output it as segmented image feature data, establish a first correspondence between the segmented image data and the segmented image feature data, input the segmented image feature data into a preset database for matching, output the matching results, establish a second correspondence between the segmented image data and the matching results, and output it as image improvement suggestion data.

2. The art design evaluation system based on data analysis as claimed in claim 1, characterized in that: The data collection and processing module includes a data collection unit and a data processing unit; The data collection unit is used to collect the current work image data, the current work text evaluation data, the current work score data set and the current artist historical auction price data set, and merge the current work image data, the current work text evaluation data, the current work score data set and the current artist historical auction price data set to output as work data; The data processing unit is used to pre-process the work data, and the pre-processing is to perform word segmentation processing, stop word removal processing and duplicate removal processing on the current work text evaluation data.

3. The art design evaluation system based on data analysis as claimed in claim 2, characterized in that: The feature extraction and analysis module includes a data analysis unit, a data calculation unit and a data integration unit; The data analysis unit is used to extract image features from the current work image data, and the image feature extraction process is as follows: Convert the current work image data to Lab color space, calculate the color histogram of the current work image data in Lab color space, obtain the pixel ratio of each color, output as a color ratio array, calculate the mean, and the mathematical expression is: Among them, μ is the mean, p k is the kth color ratio, k is an integer greater than 1, and n is the total number of non-zero color intervals; Calculate the standard deviation, the mathematical expression is: Where, σ is the standard deviation; Calculate the maximum standard deviation, the mathematical expression is: Among them, σ max is the maximum standard deviation; Calculate the richness, the mathematical expression is: Use the edge detection algorithm to extract the edge lines of the current work image data, obtain the pixel position of the edge lines, calculate the gradient angle data, divide the gradient angle data into 18 intervals with 10 degrees as the gradient value, construct the direction histogram, and calculate the information entropy of the direction distribution. The mathematical expression is: Entropy=-∑p i log(p i ) Among them, Entropy is information entropy, p i is the probability distribution of the i-th interval in the direction histogram; Combine the richness and information entropy inputs and output them as the current image feature data; The current work text evaluation data is input into the sentiment analysis model. The logic of the sentiment analysis model is: Use SentiWordNet to perform sentiment annotation on the text evaluation data of the current work. The SentiWordNet includes positive words, scores corresponding to positive words, negative words, scores corresponding to negative words, neutral words and scores corresponding to neutral words. Obtain the scores corresponding to the positive words and the scores corresponding to the negative words of the text evaluation data of the current work, and calculate the positive weight and the negative weight respectively. The mathematical expression is: The sentiment tendency value is calculated based on the positive weight and negative weight. The mathematical expression is: Sentiment tendency value = positive weight + negative weight Output the sentiment tendency value as the current sentiment analysis data.

4. The art design evaluation system based on data analysis as claimed in claim 3, characterized in that: The data calculation unit is used to calculate the current work score mean of the current work score data set, and output it as the current work score mean data, and calculate the current artist historical auction price mean of the current artist historical auction price data set, and output it as the current artist historical auction price mean data; The data integration unit is used to combine the current image feature data, the current sentiment analysis data, the current work score average data and the current artist historical auction price average data to output as score reference data.

5. The art design evaluation system based on data analysis as claimed in claim 4, characterized in that: The scoring model design module includes a model building unit and a model output unit; The model building unit is used to build a work analysis model, and the logic of the work analysis model is: Receive image feature data, sentiment analysis data, work rating average data, and historical auction price average data; Weight the richness, information entropy, sentiment value, work rating average data, and historical auction price average data: The weight of sentiment value is 0.25, the weight of richness is 0.2, the weight of mean data of work rating is 0.2, the weight of mean data of historical auction price is 0.2, and the weight of information entropy is 0.15; Calculate the total weight of the richness, information entropy, sentiment tendency value, work rating mean data and historical auction price mean data, and output it as rating data; The model output unit is used to input the scoring reference data into the work scoring model, obtain the scoring data, and output the current work scoring data.

6. The art design evaluation system based on data analysis as claimed in claim 5, characterized in that: The case selection recommendation module includes an image segmentation unit, an image matching unit and an improved output unit; The image segmentation unit is used to obtain the current work rating data. If the current work rating data is lower than the preset average work rating data, the Mask RCNN model is used to segment the current work image data and output it as a number of segmented image data. Image feature extraction is performed on the segmented image data. The image feature extraction process is the same as that of the data analysis unit. The richness and information entropy are obtained and output as segmented image richness and segmented image information entropy. The segmented image richness and segmented image information entropy are combined and output as segmented image feature data, and a first corresponding relationship between the segmented image data and the segmented image feature data is established.

7. The art design evaluation system based on data analysis as claimed in claim 6, characterized in that: The image matching unit is used to input the segmented image feature data into a database, where the database includes case segmented image data, case scoring scores corresponding to the case segmented images, case segmented image richness and case segmented image information entropy, calculate the cosine similarity between the segmented image richness and the case segmented image richness, and output it as richness similarity data, calculate the cosine similarity between the segmented image information entropy and the case segmented image information entropy, and output it as complexity similarity data, calculate the sum of the richness similarity data and the complexity similarity data, and output it as total similarity data, select the case segmented image data whose total similarity data is greater than a preset total similarity threshold, and output it as a similar image data set as a matching result.

8. The art design evaluation system based on data analysis as claimed in claim 7, characterized in that: The improvement output unit is used to select the case segmentation image data with the highest case score in the matching results, establish a second corresponding relationship between the segmented image data and the matching results, and output it as image improvement suggestion data.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the art design evaluation system based on data analysis as described in any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the art design evaluation system based on data analysis according to any one of claims 1 to 8 is implemented.

Images