Oil painting authenticity intelligent identification method, system and equipment based on intelligent vision and storage medium

Abstract

The invention relates to the field of oil painting identification, in particular to an intelligent oil painting authenticity identification method, system and device based on intelligent vision and a storage medium. When an oil painting identification instruction is received, oil painting image data are obtained through a vision detection end; extracting target features of the oil painting image data; according to the extracted target features, performing feature fusion on the target features to generate a comprehensive feature vector; inputting the comprehensive feature vector and the oil painting image data into a preset analysis model, generating a comparison feature according to the oil painting image data through the analysis model, calculating the distribution deviation degree of the comprehensive feature vector and the comparison feature, and outputting a calculation result; when the deviation degree is smaller than a preset deviation threshold value, it is judged that the oil painting image is an authentic oil painting; and when the deviation degree is greater than or equal to a preset deviation threshold value, judging that the oil painting image is a forged oil painting. Fusion analysis of the multi-dimensional features enables authenticity identification of the oil painting to be more accurate, and occurrence of misjudgment is effectively reduced.

Description

Technical Field

[0001] This application relates to the technical field of oil painting authentication, and particularly to an intelligent authentication method, system, device, and storage medium for the authenticity of oil paintings based on intelligent vision. Background Art

[0002] The authenticity authentication of oil paintings using intelligent vision is a process of using artificial intelligence (AI) and computer vision technology to intelligently analyze the images of oil painting works to determine their authenticity. Different from traditional manual authentication methods, intelligent vision technology mainly relies on technologies such as image processing, deep learning, and machine learning to distinguish the authenticity of oil paintings from the image level and data analysis level.

[0003] Existing oil painting authenticity detection technologies usually rely on the detection of a single feature, such as pigment composition analysis, canvas material, painting style, brushstroke traces, etc. These technologies judge the authenticity of a painting by detecting whether a certain feature conforms to the standards of a historical period. Forgers usually imitate a certain feature, rather than comprehensively imitating all the details of the entire work. For example, a forger may only copy a certain specific pigment or brushstroke style, but ignore other differences (such as restoration traces or the aging changes of pigments). Therefore, the analysis of a single feature may not be able to comprehensively judge the authenticity of a painting, and may even misjudge some forged oil paintings as genuine. Secondly, current anomaly detection methods usually rely on fixed standard parameters to identify "abnormal" features in oil paintings, such as pigment composition, color change, restoration traces, etc. By comparing with historical standards, any signs that do not conform to these standards will be judged as possible forgeries. As an oil painting ages, it will undergo an aging process, and the changes in pigment composition and color offset are natural phenomena, and restoration work may also affect the appearance of the painting. Therefore, simply judging anomalies based on fixed standard parameters is likely to misjudge these natural changes and restorations as signs of forgery. This method ignores the historical accumulation of oil paintings and is difficult to accurately distinguish genuine works from forgeries.

[0004] Therefore, there are defects in the existing technology and improvements are needed. Summary of the Invention

[0005] In order to solve one or several problems in the existing technology, the main purpose of this application is to provide an intelligent authentication method, system, device, and storage medium for the authenticity of oil paintings based on intelligent vision.

[0006] To achieve the above invention purpose, this application proposes an intelligent authentication method for the authenticity of oil paintings based on intelligent vision, and the method includes:

[0007] When receiving an oil painting authentication instruction, obtain oil painting image data through the visual detection end;

[0008] Extract the target features of the oil painting image data, where the target features include color features, texture features, and structural features;

[0009] According to the extracted target features, fuse the target features to generate a comprehensive feature vector;

[0010] Input the comprehensive feature vector and the oil painting image data into a preset analysis model. The analysis model generates comparison features based on the oil painting image data, calculates the distribution deviation degree between the comprehensive feature vector and the comparison features, and outputs the calculation result;

[0011] When the deviation degree is less than the preset deviation threshold, it is determined that the oil painting image is an authentic oil painting;

[0012] When the deviation degree is greater than or equal to the preset deviation threshold, it is determined that the oil painting image is a forged oil painting.

[0013] The embodiment of the present application further provides an intelligent authenticity identification system for oil paintings based on intelligent vision, including:

[0014] An acquisition module, configured to obtain oil painting image data through the visual detection end when receiving an oil painting identification instruction;

[0015] An extraction module, configured to extract the target features of the oil painting image data, where the target features include color features, texture features, and structural features;

[0016] A fusion module, configured to fuse the target features according to the extracted target features to generate a comprehensive feature vector;

[0017] An input module, configured to input the comprehensive feature vector and the oil painting image data into a preset analysis model. The analysis model generates comparison features based on the oil painting image data, calculates the distribution deviation degree between the comprehensive feature vector and the comparison features, and outputs the calculation result;

[0018] A first judgment module, configured to determine that the oil painting image is an authentic oil painting when the deviation degree is less than the preset deviation threshold;

[0019] A second judgment module, configured to determine that the oil painting image is a forged oil painting when the deviation degree is greater than or equal to the preset deviation threshold.

[0020] The present application further provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the method described in any one of the above are implemented.

[0021] The present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method described in any one of the above are implemented.

[0022] The method, system, device and storage medium for intelligent identification of the authenticity of oil paintings based on intelligent vision in the embodiments of the present application comprehensively analyze by combining multiple features such as color, brushstrokes, and texture, avoiding the limitations brought by a single feature. Each detail of an oil painting may contain unique information, and forged works often cannot perfectly match in multiple dimensions at the same time. Therefore, the fusion analysis of multi-dimensional features makes the authenticity identification of oil paintings more accurate and effectively reduces the occurrence of misjudgments. The complex details of oil paintings are often manifested as a variety of intertwined textures and brushstroke changes. Relying solely on a single feature (such as color) for judgment may ignore other important information. Through feature fusion, this method can capture the detail differences of works at a broader level and identify the subtle differences that forged works cannot perfectly reproduce. The error of a single feature is likely to cause misjudgment, while the fusion analysis combines the judgments of multiple features, thus greatly reducing the impact caused by the failure of a single feature. The system can ensure that the final identification result is more robust and reliable based on the multiple comparisons and weighted analysis of each feature. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a schematic flowchart of a method for intelligent identification of the authenticity of oil paintings based on intelligent vision according to an embodiment of the present application;

[0024] Figure 2 is a schematic flowchart of a method for intelligent identification of the authenticity of oil paintings based on intelligent vision according to an embodiment of the present application;

[0025] Figure 3 is a schematic block diagram of the structure of a system for intelligent identification of the authenticity of oil paintings based on intelligent vision according to an embodiment of the present application;

[0026] Figure 4 is a schematic block diagram of the structure of a computer device according to an embodiment of the present application.

[0027] The implementation, functional features and advantages of the objectives of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0029] Refer to Figure 1 , in the embodiments of the present application, a method for intelligent identification of the authenticity of oil paintings based on intelligent vision is provided. The method includes:

[0030] S1. When receiving an oil painting authentication instruction, obtain oil painting image data through the visual detection end;

[0031] S2. Extract the target features of the oil painting image data, where the target features include color features, texture features, and structural features;

[0032] S3. According to the extracted target features, fuse the target features to generate a comprehensive feature vector;

[0033] S4. Input the comprehensive feature vector and the oil painting image data into a preset analysis model. The analysis model generates comparison features based on the oil painting image data, calculates the distribution deviation degree between the comprehensive feature vector and the comparison features, and outputs the calculation result;

[0034] S5. When the deviation degree is less than a preset deviation threshold, determine that the oil painting image is an authentic oil painting;

[0035] S6. When the deviation degree is greater than or equal to the preset deviation threshold, determine that the oil painting image is a forged oil painting.

[0036] As described in the above steps S1-S3, visual inspection equipment (such as high-resolution cameras, scanners, etc.) is used to capture the detailed features of oil paintings. The use of high-precision equipment helps to obtain the complex color, texture, brushstrokes and other detailed information on the surface of the oil painting. Different oil paintings may show different detailed features at the microscopic level. Through high-precision image acquisition, it can be ensured that the input data can fully display all the details of the oil painting, including color changes, texture texture, brushstroke direction, etc., to avoid analysis errors caused by image quality problems. Color features: Color is an important way of expression in oil paintings. Especially in paintings of different ages and styles, the use and transformation of colors have their own unique rules. Color feature extraction usually includes information such as color distribution, hue, saturation, brightness, etc. These features help to distinguish the works of different painters and the historical period characteristics of oil paintings. Texture features: Texture is the embodiment of the details of the surface of the oil painting, including the texture of the canvas, the thickness of the paint, the traces of brushstrokes, etc. Texture features can help identify whether the painting is authentic, because it is usually difficult to perfectly reproduce the unique texture and surface effects of oil paintings in forged oil paintings. Structural features: Structural features involve the overall layout and spatial relationships in oil painting images, including the spatial construction of objects in terms of proportion, layering, and distance. Structural features can help analyze the overall composition and detailed expression of paintings to ensure that the paintings conform to the painting rules of the historical period. The extraction of these features can provide multi-dimensional data support for the authenticity judgment of oil paintings. Color, texture, and structure complement each other and can reflect the creative style, historical period, and personality of the painter. Through the extraction of these features, the system can identify the subtle differences that may exist in forgeries. Feature fusion refers to the weighted combination of the extracted color features, texture features, and structural features to form a comprehensive feature vector. The purpose of fusing these features is to combine information from different dimensions to form a comprehensive description of the oil painting image. This is usually achieved through methods such as weighted averaging and principal component analysis (PCA) to reduce the misjudgment that may be caused by a single feature. Through feature fusion, the system can obtain a more comprehensive and accurate feature description, thereby enhancing the accuracy of the judgment. The interaction between each feature can make up for the limitations of a single feature and improve the model's ability to identify forged oil paintings.

[0037] As described in the above steps S4 - S6, the comprehensive feature vector is input into a preset analysis model (such as a machine learning model or a deep learning network), and is further processed by this model to generate comparison features. This model is trained based on a large number of oil painting datasets to learn the patterns and features of various genuine oil paintings. The output of the analysis model is the comparison features, which represent the similarity between the target image and the known genuine oil paintings. By learning the general features and rules of oil paintings from a large amount of training data, the analysis model can accurately identify the authenticity of the painting. Through the comparison features, the model can carefully capture the subtle differences in the oil painting, and thus make a judgment on whether it is a genuine work. The deviation calculation is achieved by comparing the differences between the comprehensive feature vector and the comparison features. Specifically, the system will calculate the similarity or distance (such as Euclidean distance, cosine similarity, etc.) between the two, and use this value to measure the matching degree between the two. If the deviation is small, it indicates that the feature similarity between the oil painting and the genuine work is high, and it is determined to be a genuine work; when the deviation degree is greater than or equal to the preset deviation threshold, it is determined that the oil painting image is a forged oil painting. The deviation calculation is a very effective error tolerance mechanism, which can avoid misjudging genuine works due to subtle natural aging differences. At the same time, setting the preset deviation threshold can flexibly control the accuracy and tolerance of the determination, and reduce the risk of misjudgment.

[0038] As described above, by combining multiple features such as color, brushstrokes, and texture for comprehensive analysis, the limitations brought by a single feature are avoided. Every detail of an oil painting work may contain unique information, and forged works often cannot perfectly match in multiple dimensions at the same time. Therefore, the fusion analysis of multi - dimensional features makes the authenticity identification of oil paintings more accurate and effectively reduces the occurrence of misjudgments. The complex details of oil paintings often manifest as various intertwined textures and brushstroke changes. Relying solely on a single feature (such as color) for judgment may ignore other important information. Through feature fusion, this method can capture the detail differences of the work at a broader level and identify the subtle differences that forged works cannot perfectly reproduce. The error of a single feature is likely to cause misjudgment, while the fusion analysis combines the judgments of multiple features, thus greatly reducing the impact caused by the failure of a single feature. The system can ensure that the final identification result is more robust and reliable based on the multiple comparisons and weighted analysis of each feature.

[0039] Refer to Figure 2 , in one embodiment, before the step of fusing the target features to generate a comprehensive feature vector, the method further includes:

[0040] S31. Analyze whether there are abnormal features in the target features according to the extracted target features;

[0041] S32. When there is no abnormal feature in the target feature, the target feature meets the condition for generating a comprehensive feature vector;

[0042] S33. Based on the judgment result, perform feature fusion on the target feature to generate a comprehensive feature vector.

[0043] As described in the above steps, in the process of authenticating oil paintings, it is first necessary to extract target features. These features can be the color, texture, shape, brushstrokes, etc. of the image. The process of extracting these target features is based on computer vision technology and uses image processing algorithms (such as edge detection, texture analysis, etc.) to extract representative and distinguishable features from the image. These features can help the system capture the detailed differences in oil painting works and provide basic data for subsequent analysis. The extracted target features provide an accurate basis for subsequent analysis, ensuring that the subsequent steps can rely on accurate data for judgment and fusion. Preparing for feature fusion by obtaining detailed information from various angles (such as color, brushstrokes, etc.) improves the comprehensiveness of oil painting authenticity identification. The step of target feature anomaly detection is mainly used to determine whether there are unreasonable or abnormal features among the extracted features. Common abnormal features include irregular color distributions in the image, prominent texture deviations, etc. The existence of abnormal features may indicate that the work is inconsistent with real oil painting works, and these abnormalities can usually be detected by setting thresholds. For example, if a certain feature deviates too much from the features of most oil painting works, then this feature may be regarded as abnormal. The role of anomaly detection is to identify potential forgery features at an early stage and avoid wrongly including works that do not conform to the norm in the authentication process. This method increases the robustness of the system, excludes untrustworthy target features before feature fusion, and ensures that the final authentication result is more accurate. This step can also effectively reduce misjudgments caused by single feature anomalies and improve the system's ability to identify real oil painting works. After confirming that there are no anomalies in the target features, it is judged whether feature fusion can continue. Only when the target features fall within the common feature range of oil painting works is it allowed to continue with feature fusion. This process determines whether the target features meet the conditions for generating a comprehensive feature vector through some form of judgment criteria or rules. For example, if the distribution and form of the target features match those of known oil painting works, then the next step of fusion can be carried out. Otherwise, the system may determine that the features are invalid or unrepresentative. This step ensures that only target features that meet certain criteria are used to generate a comprehensive feature vector, which can effectively improve the accuracy and effectiveness of subsequent analysis. Through this condition screening mechanism, the influence of invalid or greatly deviated features on the final result is avoided, ensuring that the fused comprehensive feature vector represents the true features of the oil painting work. Once the target features meet the conditions, the next step is to perform feature fusion, that is, to weight and integrate each target feature to generate a comprehensive feature vector. The fusion process usually takes into account the relative importance of different features and, based on mathematical methods such as weighted algorithms, PCA (principal component analysis), etc., maximally retains the effective information. The purpose of feature fusion is to integrate multiple independent features into a unified feature representation, thereby enhancing the accurate description of the overall features of oil painting works.The comprehensive feature vector after feature fusion better embodies the overall characteristics of the oil painting work and can accurately reflect the subtle differences between genuine and forged works. This comprehensive feature vector provides a more comprehensive and reliable basis for subsequent judgment, classification, and analysis, thereby improving the effect of oil painting authenticity identification. Due to the higher integration degree of the fused features, the system can quickly and accurately make identification decisions in complex situations.

[0044] In one embodiment, the steps of analyzing whether there are abnormal features in the target features include:

[0045] Obtain the historical background information of the oil painting, where the background parameters include the creation time of the oil painting, restoration history parameters, and material parameters;

[0046] According to the historical background information, construct a time analysis model for inferring the historical change matching degree between the target features and the background parameters based on timestamps;

[0047] Obtain the color features, texture features, and structural features of the target features;

[0048] Respectively input the color features, texture features, and structural features into the time analysis model, and through the time analysis model, respectively compare and analyze the color features, texture features, and structural features with the historical background information, and output the historical change matching degree between the target features and the background parameters;

[0049] Based on the output matching degree, quantify the matching degree to obtain the matching coefficients of the color features, texture features, and structural features;

[0050] When the matching coefficients of the color features, texture features, and structural features are respectively greater than the preset coefficient thresholds, it is determined that there are no abnormal features in the target features;

[0051] When the matching coefficient of any one group of the color features, texture features, and structural features is less than the preset coefficient threshold, it is determined that there are abnormal features in the target features.

[0052] As described above, obtaining the historical background information of an oil painting provides background data support for analyzing whether there are abnormalities in the target features. The historical background information includes the creation time, restoration history, and materials used in the oil painting. Such information can help the system understand the historical background of the oil painting and the evolution process of its features. Creation time: It affects the artistic style, pigment formula, and technical techniques of the oil painting. For example, oil paintings in the 16th century may use materials and techniques that are completely different from those in the 20th century. Restoration history: The restoration history reflects the possible changes that the oil painting may have undergone during the historical process. For example, restoration may involve adding new pigments or changing the original colors and textures. Material parameters: The material characteristics of pigments, canvases, etc. used in different historical periods may vary, and these will all affect the surface characteristics of the oil painting. Obtaining this background information helps to understand the formation and change process of the oil painting and assists in subsequent analysis of whether the features of the oil painting conform to the expected historical trajectory. For example, the use of certain pigments or color changes can help determine whether there are anomalies in the oil painting with inconsistent time. The core goal of the time analysis model is to infer the matching degree between the target features and the historical background parameters based on the time stamps. The input of this model is the historical background information of the oil painting (including the creation time, restoration history, and material parameters), and the output is the matching degree of the target features with the historical changes of the background parameters. The basic logic of this model is to predict the color, texture, and structural features that the oil painting should have at a specific time point by analyzing factors such as the creation time, restoration process, and material changes of the oil painting. The time analysis model can deduce the expected changes of the target features based on historical data (such as the material change records during the restoration of the oil painting, the historical usage records of pigments, etc.). Through this model, the characteristic changes that the oil painting should exhibit in different historical periods can be accurately speculated. This helps to judge whether the existing features conform to the expected historical evolution trajectory, thereby identifying abnormal features that do not conform to the historical background or restoration history. If the changes in the target features exceed the range predicted by the model, it may mean that there are anomalies in these features during the historical evolution. Color features: Refer to the color distribution, color intensity, hue, etc. on the surface of the oil painting. Color features can reflect the pigment usage of the oil painting and the color trends popular during the oil painting creation period. Texture features: Refer to the fineness of the surface of the oil painting, the performance of brushstrokes, the picture hierarchy, etc. Texture features are closely related to the creation techniques of the oil painting and the painting styles of the painters. Structural features: Refer to the structural features in aspects such as the overall composition and painting techniques of the oil painting. Structural features can reveal the composition arrangement, painting techniques, and styles of the oil painting. These features are extracted from the image through image processing techniques (such as color space conversion, texture analysis, morphological analysis, etc.). By extracting these features, the system can capture the changes in the oil painting in terms of details and overall performance, thereby helping to judge whether there are obvious historical inconsistencies (such as color distribution and texture details that do not conform to the creation period). The color features, texture features, and structural features are respectively input into the time analysis model. The purpose is to separately analyze each type of feature through the model and evaluate its matching degree with the historical background parameters.Color feature comparison: Based on the creation time of the oil painting and the history of paint usage, the model can predict the color features that should appear. For example, a certain color may be common or obsolete during a certain period, so the change in color features can help determine whether it is reasonable. Texture feature comparison: Through a time model, infer the changes in brushstrokes and materials during the creation and restoration of the oil painting to help determine whether the texture conforms to the historical background. Structural feature comparison: The model can predict the composition method of the oil painting based on the historical background and evaluate whether the structural features conform to the artistic style of a specific period. By analyzing each feature independently, it is possible to more accurately determine whether the target features in each aspect conform to the historical change trend. If a certain feature (such as color or texture) does not match the historical background, possible abnormal features can be identified. The output matching degree is a quantitative index of the matching degree obtained by comparing the target features with the historical background information. This matching degree reflects the degree of conformity between the target features and the historical background deduced by the time analysis model. The output of the matching degree is usually in numerical form, such as a ratio or a fraction, indicating the matching degree between the target features and the historical background parameters. The output of the matching degree helps the system quantify the historical rationality of the target features. If the matching degree is high, it means that the target features are consistent with the creation and restoration history of the oil painting; if the matching degree is low, there may be an anomaly, indicating that the target features need further investigation. Quantifying the matching degree is achieved by calculating a matching coefficient for each feature (color, texture, structure) and comparing it with a preset threshold. This matching coefficient is usually obtained by converting the difference in features into a numerical value through an algorithm (for example, a value between 0 and 1). These coefficients reflect the consistency of each feature with the historical background information. The higher the coefficient, the better the matching degree, and the more the feature conforms to the historical expectation. If the matching coefficients of the color, texture, and structural features are all greater than the preset threshold, it is determined that the target features are normal, indicating that the oil painting features are consistent with the historical background. If the matching coefficient of any feature is lower than the threshold, it is considered that there is an anomaly in this feature, suggesting that the oil painting may contain elements that do not conform to the historical background (such as signs of improper restoration or forgery). This judgment step is the key to the entire anomaly detection process, which determines whether further review of the authenticity of the oil painting is required.

[0053] In one embodiment, after the step of determining that there is an abnormal feature in the target feature when the matching coefficient of any one of the color feature, texture feature, and structural feature is less than the preset coefficient threshold, the method further includes:

[0054] When there is an abnormal feature in the target feature, verify whether the target feature meets the conditions of forgery;

[0055] When the target feature meets the conditions of forgery, it is determined that the oil painting image is a forged oil painting.

[0056] As described above, when the matching coefficients of color, texture, and structural features are lower than the threshold, it indicates that the image has abnormal features, which provides a basis for the next step of determining whether the image is forged. The further step is to verify whether these abnormal features meet the conditions of forgery. The specific conditions may include excessive smoothing, repeated patterns, unnatural edge processing, etc. in certain local areas of the image. Forged oil paintings are often generated by certain computer algorithms, and these algorithms are difficult to replicate the details and irregularities in real oil paintings. Therefore, when the image features do not conform to the normal laws of oil paintings in certain dimensions, further verification can be carried out to confirm whether it is forged. If it is found through verification that the features of the image meet the conditions of forgery (such as excessive smoothing, obvious digital traces, etc.), then it can be finally determined that the image is a forged oil painting. Through the comprehensive analysis of color, texture, and structural features, the goal is to use the inherent visual laws of oil painting images to determine whether the image is forged. The detection of abnormal features and the verification of forgery conditions are multi-level and multi-angle analysis methods, aiming to accurately identify forged oil paintings and avoid misjudgment. Setting a threshold to determine whether abnormal features appear helps to improve the accuracy and reliability of the system.

[0057] In one embodiment, the steps of verifying whether the target features meet the conditions of forgery include:

[0058] Identify the style type of the oil painting image data;

[0059] According to the style type, obtain the style verification data and weight coefficients corresponding to the abnormal features;

[0060] Analyze the style coincidence degree between the abnormal features and the style verification data;

[0061] According to the weight coefficients, perform a weighted calculation on the style coincidence degree to obtain a weighted coincidence degree;

[0062] Quantify the weighted coincidence degree to obtain a style coincidence value;

[0063] Judge whether the style coincidence value is less than a preset coincidence threshold;

[0064] When the style coincidence value is less than the preset coincidence threshold, it is determined that the target features meet the conditions of forgery, and the oil painting image is determined to be a forged oil painting.

[0065] As mentioned above, each oil painting has its unique artistic style. Style types generally refer to the characteristics of an artist or an art movement, such as Impressionism, Realism, Abstractism, etc. Identifying the style type of an oil painting is the basis for classifying the image, and the style type will affect the color distribution, texture details, brushstroke patterns, etc. that appear in the image. Oil paintings of different styles have different expressive characteristics. For example, Impressionist oil paintings have bright colors and blurred boundaries, while Realist oil paintings focus on details and contrast between light and dark. Therefore, determining the style type of an oil painting helps to better understand the characteristics of the image and conduct subsequent feature comparison and verification. Style verification data refers to the statistical or model data related to a specific style type, usually including the common color characteristics, texture characteristics, composition, etc. in that style. The weight coefficient represents the importance or priority of these characteristics in the style. For example, some styles may pay more attention to the use of colors, while other styles may focus more on texture details. The weight coefficient helps to weight the influence of each characteristic and avoid certain unimportant characteristics having too much influence on the judgment result. Oil paintings of different styles have differences in the performance of different characteristics, so it is necessary to obtain corresponding feature data according to the style type. At the same time, by assigning different weight coefficients to different characteristics, the degree of style conformity of the image can be evaluated more precisely, avoiding the influence of certain irrelevant characteristics on the final judgment. The style coincidence degree refers to the consistency between the abnormal features extracted from the oil painting image and the specific style verification data. By calculating the similarity between the abnormal features and the style verification data, the degree of conformity of the style features shown in the image with the normal oil painting style can be quantified. If the abnormal features are highly consistent with the style verification data, it indicates that the image may be a genuine oil painting; if the coincidence degree is low, there may be a risk of forgery. Abnormal features indicate the inconsistency of the image in some aspects with the standard oil painting works, which may be a sign of a forged image. By calculating the style coincidence degree, the abnormal features can be compared with the real style data to accurately evaluate the matching situation of the style. Analyzing the style coincidence degree can help to determine whether the performance of the abnormal features in terms of style meets the expectations. Weighted calculation is the process of adjusting the style coincidence degree through the weight coefficient. Different style features may have different importance in certain styles. For example, in the Realist style, texture and details may be more important, while in the Impressionist style, the coincidence degree of colors may be more crucial. Through weighted calculation, the system can comprehensively evaluate the overall style conformity degree according to the importance of each characteristic. The importance of characteristics is different under different styles, and some characteristics may be more crucial for the expression of the style than others. By assigning different weight coefficients to these characteristics, it is possible to more accurately evaluate whether the image conforms to the style and avoid certain characteristics affecting the judgment. Through weighted calculation, a comprehensive style coincidence degree can be obtained according to the importance of different characteristics, thereby improving the accuracy of the judgment. The style coincidence value is the quantified result of the weighted coincidence degree, usually represented by a numerical value, reflecting the matching degree between the image style and the expected style.The larger the value, the more the image style conforms to the expected style; the smaller the value, the greater the difference between the image style and the expected style. The quantified style coincidence value makes the analysis results more operable and comparable. By quantifying the coincidence degree, clear numerical support can be provided for subsequent judgments, reducing the interference of subjective human judgment. Through the style coincidence value, the system can more clearly judge whether the image conforms to the expected style, providing a quantitative basis for subsequent judgments on forgery. The preset coincidence threshold is the standard for judging whether the style coincidence degree is sufficiently in line with the expected style. If the style coincidence value is lower than this threshold, it indicates that the matching degree of the image in terms of style features is insufficient, and the image may be a forged one; if the style coincidence value is higher than the threshold, the image is considered to conform to this style and belongs to a genuine oil painting. By setting a standard threshold, it is possible to effectively judge whether the style coincidence degree is high enough to avoid misjudgment. The setting of the threshold can be adjusted according to the actual situation to improve the accuracy of the system. This step can judge whether the image conforms to the style of a genuine oil painting according to the set standard, so as to determine whether the image is forged. If the style coincidence value is lower than the preset threshold, it indicates that there is a large difference between the image style and the genuine oil painting style, and the image may be a forged oil painting. Through this judgment, the system determines whether the image is a forged oil painting. This step is the final judgment basis. By comprehensively analyzing the style features of the image and the verification data, it is judged whether the forgery conditions are met, avoiding misjudgment of forged oil paintings. Through this final step, the system can accurately judge the image and confirm whether it is a forged oil painting.

[0066] In one embodiment, after the step of judging whether the style coincidence value is less than the preset coincidence threshold, the method further includes:

[0067] When the style coincidence value is greater than or equal to the preset coincidence threshold, it is judged that the target feature does not meet the forgery condition;

[0068] Based on the judgment result, the target features are fused to generate a comprehensive feature vector.

[0069] As described above, even if some features of the image do not directly show signs of forgery, it is still necessary to conduct a multi-dimensional comprehensive analysis of the overall style, structure, texture, etc. of the image through the integrated feature vector. This means that the system will fuse various features of the image into a high-dimensional feature vector to further analyze the authenticity of the image. Sometimes, a single feature may not be sufficient to fully reveal whether the image is forged, especially when the forged image closely mimics the real work in some aspects. By generating an integrated feature vector, detailed analysis can be carried out from multiple dimensions to capture signs of forgery that may be overlooked by a single feature. In this way, even if the target feature does not clearly show signs of forgery, the system can still ensure a comprehensive determination by integrating multiple features. As a representation containing multi-dimensional information of the image, the integrated feature vector can effectively complement what may be missed by individual features, making the final analysis more accurate.

[0070] In one embodiment, the style types of the oil painting image data include Impressionism, Baroque, and Abstract. Impressionism focuses on the instantaneous feeling and the light and shadow changes in nature. Baroque expresses emotions and dynamics through dramatic light and shadow effects and complex compositions. Abstract art breaks the traditional forms of expression and explores the beauty of shapes and colors themselves. Impressionism tends to use brighter and more scattered colors, Baroque uses strong light and dark contrasts, and Abstract art mainly features color blocks and pure color combinations. The brushstrokes of Impressionism are obvious, those of Baroque are more delicate, and Abstract art may show more free and irregular brushstrokes. The composition of Baroque is complex, paying attention to details and dynamics; Abstract art has almost no traditional composition and emphasizes the free combination of forms.

[0071] Refer to Figure 3 , and an intelligent authentication system for the authenticity of oil paintings based on intelligent vision is also provided in the embodiment of the present application, including:

[0072] An acquisition module 1, configured to obtain oil painting image data through the visual detection end when receiving an oil painting authentication instruction;

[0073] An extraction module 2, configured to extract target features of the oil painting image data, where the target features include color features, texture features, and structural features;

[0074] A fusion module 3, configured to fuse the target features according to the extracted target features to generate an integrated feature vector;

[0075] An input module 4, configured to input the integrated feature vector and the oil painting image data into a preset analysis model, generate comparison features according to the oil painting image data through the analysis model, calculate the distribution deviation degree between the integrated feature vector and the comparison features, and output the calculation result;

[0076] The first judgment module 5 is configured to judge that the oil painting image is an authentic oil painting when the deviation degree is less than a preset deviation threshold.

[0077] The second judgment module 6 is configured to judge that the oil painting image is a forged oil painting when the deviation degree is greater than or equal to the preset deviation threshold.

[0078] As described above, it can be understood that each component of the intelligent authentication system for the authenticity of oil paintings based on intelligent vision proposed in this application can implement the functions of any one of the above-described intelligent authentication methods for the authenticity of oil paintings based on intelligent vision, and the specific structure will not be elaborated.

[0079] Referring to Figure 4 , an embodiment of this application also provides a computer device, which may be a server, and its internal structure may be as Figure 4 shown. The computer device includes a processor, a memory, a network interface, and a database connected through a system bus. Among them, the processor of the computer design is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data such as monitoring data. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements an intelligent authentication method for the authenticity of oil paintings based on intelligent vision.

[0080] The above processor executes the above-mentioned intelligent authentication method for the authenticity of oil paintings based on intelligent vision, including: when receiving an oil painting authentication instruction, obtaining oil painting image data through the visual detection end; extracting target features of the oil painting image data, where the target features include color features, texture features, and structural features; according to the extracted target features, performing feature fusion on the target features to generate a comprehensive feature vector; inputting the comprehensive feature vector and the oil painting image data into a preset analysis model, generating a comparison feature by the analysis model according to the oil painting image data, calculating the distribution deviation degree between the comprehensive feature vector and the comparison feature, and outputting the calculated result; when the deviation degree is less than a preset deviation threshold, judging that the oil painting image is an authentic oil painting; when the deviation degree is greater than or equal to the preset deviation threshold, judging that the oil painting image is a forged oil painting.

[0081] An embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements an intelligent authenticity identification method for oil paintings based on intelligent vision, including the steps of: when receiving an oil painting identification instruction, obtaining oil painting image data through the vision detection end; extracting target features of the oil painting image data, where the target features include color features, texture features, and structural features; according to the extracted target features, fusing the target features to generate a comprehensive feature vector; inputting the comprehensive feature vector and the oil painting image data into a preset analysis model, generating a comparison feature by the analysis model according to the oil painting image data, calculating the distribution deviation degree between the comprehensive feature vector and the comparison feature, and outputting the calculated result; when the deviation degree is less than a preset deviation threshold, determining that the oil painting image is an authentic oil painting; when the deviation degree is greater than or equal to the preset deviation threshold, determining that the oil painting image is a forged oil painting.

[0082] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above method embodiments. Among them, any reference to a memory, storage, database, or other medium provided in the present application and used in the embodiments can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0083] It should be noted that in this text, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article or method comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or elements inherent to such process, apparatus, article or method. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, apparatus, article or method comprising such element.

[0084] The above are only the preferred embodiments of the present application, and do not limit the patent scope of the present application accordingly. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present application, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present application.

Claims

1. An intelligent method for identifying the authenticity of oil paintings based on intelligent vision, characterized in that: The method comprises: When receiving an oil painting identification instruction, acquiring oil painting image data through the visual detection terminal; Extracting target features of the oil painting image data, wherein the target features include color features, texture features and structural features; According to the extracted target features, the target features are subjected to feature fusion to generate a comprehensive feature vector; Input the comprehensive feature vector and the oil painting image data into a preset analysis model, generate contrast features according to the oil painting image data through the analysis model, calculate the distribution deviation degree of the comprehensive feature vector and the contrast features, and output the calculation result; When the degree of deviation is less than a preset deviation threshold, the oil painting image is judged to be an authentic oil painting; When the degree of deviation is greater than or equal to a preset deviation threshold, the oil painting image is determined to be a forged oil painting.

2. The intelligent identification method of oil painting based on intelligent vision according to claim 1 is characterized in that: The target features are subjected to feature fusion, Before the step of generating the comprehensive feature vector, the method further comprises: Analyzing whether the target features have abnormal features according to the extracted target features; When the target feature does not have abnormal features, the target feature meets the conditions for generating a comprehensive feature vector; Based on the judgment result, the target features are fused to generate a comprehensive feature vector.

3. The intelligent identification method of oil painting based on intelligent vision according to claim 2 is characterized in that: The step of analyzing whether the target feature has an abnormal feature comprises: Acquiring historical background information of the oil painting, wherein the background parameters include the creation time, restoration history parameters and material parameters of the oil painting; Based on the historical background information, a time analysis model is constructed to infer the degree of matching between the historical changes of the target features and the background parameters according to the timestamp; Acquire color features, texture features and structural features of the target features; Inputting the color features, texture features and structure features into the time analysis model respectively, comparing and analyzing the color features, texture features and structure features with historical background information respectively through the time analysis model, and outputting the matching degree of historical changes of target features and background parameters; Based on the output matching degree, quantify the matching degree to obtain the matching coefficients of the color feature, texture feature and structure feature; When the matching coefficients of the color feature, texture feature and structure feature are respectively greater than the preset coefficient thresholds, it is determined that the target feature does not have abnormal features; When the matching coefficient of any group of the color feature, texture feature and structural feature is less than a preset coefficient threshold, it is determined that the target feature has an abnormal feature.

4. The intelligent identification method of oil painting based on intelligent vision according to claim 3 is characterized in that: After the step of determining that the target feature has an abnormal feature when the matching coefficient of any one of the color feature, texture feature and structural feature is less than a preset coefficient threshold, the method further includes: When the target feature has an abnormal feature, verify whether the target feature meets the forgery condition; When the target feature meets the forgery condition, the oil painting image is determined to be a forged oil painting.

5. The intelligent method for identifying the authenticity of oil paintings based on intelligent vision according to claim 4 is characterized in that: The step of verifying whether the target feature meets the forgery condition comprises: Identifying the style type of the oil painting image data; According to the style type, obtaining style verification data and weight coefficient corresponding to the abnormal feature; Analyzing the style overlap between the abnormal feature and the style verification data; According to the weight coefficient, weighted calculation is performed on the style coincidence degree to obtain a weighted coincidence degree; quantifying the weighted coincidence to obtain a style coincidence value; Determining whether the style overlap value is less than a preset overlap threshold; When the style coincidence value is less than a preset coincidence threshold, it is determined that the target feature meets the forgery condition, and the oil painting image is determined to be a forged oil painting.

6. The intelligent identification method of oil painting based on intelligent vision according to claim 5 is characterized in that: After the step of determining whether the style overlap value is less than a preset overlap threshold, the method further includes: When the style overlap value is greater than or equal to a preset overlap threshold, it is determined that the target feature does not meet the forgery condition; Based on the judgment result, the target features are fused to generate a comprehensive feature vector.

7. The intelligent method for identifying the authenticity of oil paintings based on intelligent vision according to claim 5, characterized in that: The style types of the oil painting image data include impressionism, baroque and abstraction.

8. An intelligent oil painting authenticity identification system based on intelligent vision, characterized in that: include: An acquisition module, used for acquiring oil painting image data through the visual detection end when receiving an oil painting identification instruction; An extraction module, used for extracting target features of the oil painting image data, wherein the target features include color features, texture features and structural features; A fusion module, used to fuse the target features according to the extracted target features to generate a comprehensive feature vector; An input module, used to input the comprehensive feature vector and the oil painting image data into a preset analysis model, generate contrast features according to the oil painting image data through the analysis model, calculate the distribution deviation degree between the comprehensive feature vector and the contrast features, and output the calculation result; A first judgment module, configured to judge that the oil painting image is an authentic oil painting when the deviation degree is less than a preset deviation threshold; The second judgment module is used to judge that the oil painting image is a forged oil painting when the deviation degree is greater than or equal to a preset deviation threshold.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images