Cultural relic feature processing method and device, electronic equipment and storage medium

Abstract

The invention provides a cultural relic feature processing method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the image segmentation processing of a cultural relic image of a to-be-analyzed cultural relic, and obtaining an image block set of the cultural relic image; determining significance evaluation indexes of the image blocks in the image block set by applying a predetermined standard evaluation index of a category to which the to-be-analyzed cultural relic belongs; wherein the significance evaluation index represents the richness degree of information contained in the image block in the cultural relic image; screening out a feature block set of the cultural relic image from the image block set according to significance evaluation indexes of the image blocks in the image block set; and determining feature information of the cultural relic image according to the feature block set. Through the technology disclosed by the invention, the accuracy and efficiency of extracting the salient features on the surface of the cultural relic can be improved, and deeper and more comprehensive cultural relic feature data are provided for the fields of cultural relic protection, research, identification and the like.

Description

Technical Field

[0001] The present disclosure relates to the field of image processing technologies, and particularly to a method, apparatus, electronic device, and storage medium for processing cultural relic features. Background Art

[0002] In the fields of cultural heritage protection and archaeological research, it is an important task to conduct a detailed analysis of the features on the surface of cultural relics, especially the microscopic texture features. The microscopic texture on the surface of cultural relics not only contains rich historical information but also is the key information for understanding its manufacturing process and material properties.

[0003] Traditional methods for analyzing the surface of cultural relics mainly rely on the visual inspection and manual analysis by experts. This method is not only time-consuming and laborious but also has limitations in terms of the accuracy and efficiency of the analysis. Summary of the Invention

[0004] The purpose of the present disclosure is to provide a method, apparatus, electronic device, and storage medium for processing cultural relic features, so as to improve the accuracy and efficiency of extracting the significant features on the surface of cultural relics and provide more in-depth and comprehensive cultural relic feature data for fields such as cultural relic protection, research, and identification.

[0005] An embodiment of the present disclosure provides a method for processing cultural relic features. The method includes: performing image segmentation processing on the cultural relic image of the cultural relic to be analyzed to obtain an image block set of the cultural relic image; applying a standard evaluation index of the category to which the cultural relic to be analyzed belongs, which is determined in advance, to determine a significant evaluation index of the image blocks in the image block set; where the significant evaluation index represents the richness of the information contained in the image block in the cultural relic image; screening out a feature block set of the cultural relic image from the image block set according to the significant evaluation index of the image blocks in the image block set; and determining the feature information of the cultural relic image according to the feature block set.

[0006] In a second aspect, an embodiment of the present disclosure further provides a device for processing cultural relic features. The device includes: a segmentation processing module for performing image segmentation processing on the cultural relic image of the cultural relic to be analyzed to obtain an image block set of the cultural relic image; an evaluation index determination module for applying a standard evaluation index of the category to which the cultural relic to be analyzed belongs, which is determined in advance, to determine a significant evaluation index of the image blocks in the image block set; where the significant evaluation index represents the richness of the information contained in the image block in the cultural relic image; a block screening module for screening out a feature block set of the cultural relic image from the image block set according to the significant evaluation index of the image blocks in the image block set; and a feature information determination module for determining the feature information of the cultural relic image according to the feature block set.

[0007] In a third aspect, an embodiment of the present disclosure further provides an electronic device, including a processor and a memory. The memory stores computer-executable instructions that can be executed by the processor. The processor executes the computer-executable instructions to implement the above-mentioned method for processing cultural relic features.

[0008] In a fourth aspect, an embodiment of the present disclosure further provides a computer-readable storage medium. The computer-readable storage medium stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions cause the processor to implement the above-mentioned method for processing cultural relic features.

[0009] An embodiment of the present disclosure provides a method, device, electronic device, and storage medium for processing cultural relic features. By performing image segmentation processing on the cultural relic image of the cultural relic to be analyzed, the cultural relic image can be divided into multiple smaller-sized image blocks, such that the information content and quantity included in each image block are different. On this basis, applying the pre-determined standard evaluation index of the category to which the cultural relic to be analyzed belongs, the significance evaluation index of each image block is determined. The significance evaluation index characterizes the richness of the information contained in the image block in the cultural relic image, so as to quantitatively measure each image block. Furthermore, according to the significance evaluation index of the image block, a set of feature blocks is selected from the set of image blocks. The feature blocks in the set of feature blocks contain relatively rich information. Therefore, the feature blocks in the set of feature blocks can reflect the detailed information of the cultural relic image, and the set of feature blocks removes the non-representative image blocks in the set of image blocks. While improving the accuracy of extracting the significant features on the surface of the cultural relic, it can also reduce the computational amount of determining the feature information of the above-mentioned cultural relic image, improve the efficiency of feature extraction, and provide more in-depth and comprehensive cultural relic feature data for fields such as cultural relic protection, research, and identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the specific embodiments of the present disclosure or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. Obviously, the following drawings are some embodiments of the present disclosure. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0011] Figure 1 It is a flowchart of a method for processing cultural relic features provided by an embodiment of the present disclosure;

[0012] Figure 2 It is a schematic diagram of a partial image of a calligraphy and painting cultural relic provided by an embodiment of the present disclosure;

[0013] Figure 3 It is a schematic diagram after segmentation of a partial image of a calligraphy and painting cultural relic provided by an embodiment of the present disclosure;

[0014] Figure 4 Schematic diagram of evaluation indexes for image segmentation of a partial image of a calligraphy and painting cultural relic provided by an embodiment of the present disclosure;

[0015] Figure 5 Schematic diagram of a characteristic region of a partial image of a calligraphy and painting cultural relic provided by an embodiment of the present disclosure;

[0016] Figure 6 Schematic diagram of characteristic points of a partial image of a calligraphy and painting cultural relic provided by an embodiment of the present disclosure;

[0017] Figure 7 Schematic diagram of the structure of a processing device for cultural relic characteristics provided by an embodiment of the present disclosure;

[0018] Figure 8 Schematic diagram of the structure of an electronic device provided by an embodiment of the present disclosure. Specific embodiments

[0019] The technical solutions of the present disclosure will be described clearly and completely below in conjunction with the embodiments. Obviously, the described embodiments are some, but not all, of the embodiments of the present disclosure. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present disclosure without creative efforts shall fall within the protection scope of the present disclosure.

[0020] For cultural relics such as calligraphy, seals, paintings, and pottery, their surface textures contain a large amount of information, which not only contains rich historical information but also is the key to understanding their manufacturing processes and material properties. If the analysis of the surface of cultural relics mainly relies on the visual inspection and manual analysis of experts, it is time-consuming and laborious, and there are also limitations in terms of the accuracy and efficiency of the analysis. If the analysis of the surface of cultural relics mainly relies on traditional image processing techniques, such as various filters, edge detection algorithms, and image enhancement techniques to identify and analyze surface textures, the accuracy is poor, and it often requires intervention by people with relatively high professional knowledge. Based on this, the embodiments of the present disclosure provide a method, device, electronic device, and storage medium for processing cultural relic characteristics to improve the accuracy and efficiency of cultural relic characteristic processing.

[0021] See Figure 1 The flowchart of a method for processing cultural relic characteristics shown. This method can be applied to an electronic device, which can be a server on the network side or a terminal device on the user side. The method includes the following steps:

[0022] Step S102, perform image segmentation processing on the cultural relic image of the cultural relic to be analyzed to obtain a set of image segments of the cultural relic image.

[0023] The above image segmentation process can divide the cultural relic image into multiple image patches, and the shape and size of the image patches can be determined according to the type of cultural relics. For example, the image patches of calligraphy can be slightly larger in size than those of seals and paintings.

[0024] During the image segmentation process, a certain number of large image patches can be first divided, and then the large image patches can be segmented to obtain smaller-sized image patches.

[0025] The cultural relic image of the cultural relic to be analyzed can be an image collected by a high-resolution image acquisition device, or an image obtained by performing preliminary analysis and preprocessing on the collected image to remove the environmental interference area.

[0026] Step S104: Apply the standard evaluation index of the category to which the cultural relic to be analyzed belongs that has been predetermined to determine the significance evaluation index of the image patches in the image patch set; wherein, the significance evaluation index represents the richness of the information contained in the image patch in the cultural relic image.

[0027] The above standard evaluation index can be artificially predetermined in advance according to the common characteristics of the images of the category to which the cultural relic to be analyzed belongs, or by collecting the images of the category to which the cultural relic to be analyzed belongs and submitting them to an AI model for analysis to determine the standard evaluation index of the category to which the cultural relic to be analyzed belongs. This standard evaluation index can be determined based on multiple factors, such as texture complexity, color change, edge information, etc. The standard evaluation index is used to stipulate that the significance evaluation index of the image patch corresponding to the key area worthy of further analysis in the image is higher than that of the image patch corresponding to the non-key area.

[0028] The magnitude of the significance evaluation index can reflect the richness of the information contained in the image patch in the entire cultural relic image, and this richness can include the importance and uniqueness degree of the image patch in the entire cultural relic image. The significance evaluation index is positively correlated with the above richness, that is, the higher the significance evaluation index of the image patch, the higher the richness of the information contained in the image patch in the entire cultural relic image; vice versa; the significance evaluation index can be set as a value between 0 and 1.

[0029] Step S106: According to the significance evaluation index of the image patches in the above image patch set, screen out the feature patch set of the above cultural relic image from the image patch set.

[0030] Based on the fact that the significance evaluation index of the above image patch represents the richness of the information contained in the image patch in the cultural relic image, therefore, the image patches with high significance evaluation indexes contain a large amount of information, and these image patches with high significance evaluation indexes can be screened out as the feature patches of the cultural relic image.

[0031] Step S108: Determine the feature information of the cultural relic image according to the above-mentioned feature block set.

[0032] The feature blocks in the above-mentioned feature block set are selected according to the significance evaluation index of image blocks. Therefore, the feature blocks in the feature block set can reflect the detailed information of the cultural relic image, and the feature block set removes the unrepresentative image blocks in the image block set, which can reduce the calculation amount of determining the feature information of the above-mentioned cultural relic image while ensuring the accuracy of the feature information.

[0033] By performing image segmentation processing on the cultural relic image of the cultural relic to be analyzed, the above method can divide the cultural relic image into multiple smaller-sized image blocks, so that the information content and quantity included in each image block are different. On this basis, the pre-determined standard evaluation index of the category to which the cultural relic to be analyzed belongs is applied to determine the significance evaluation index of each image block, and the significance evaluation index is used to characterize the richness of the information contained in the image block in the cultural relic image, so as to quantitatively measure each image block. Furthermore, according to the significance evaluation index of the image block, a feature block set is selected from the image block set. The feature blocks in the feature block set contain relatively rich information. Therefore, the feature blocks in the feature block set can reflect the detailed information of the cultural relic image, and the feature block set removes the unrepresentative image blocks in the image block set, which can improve the accuracy of the extraction of the significant features on the surface of the cultural relic while reducing the calculation amount of determining the feature information of the above-mentioned cultural relic image, improving the efficiency of feature extraction, and providing more in-depth and comprehensive cultural relic feature data for fields such as cultural relic protection, research, and identification.

[0034] As a possible implementation manner, the above-mentioned image segmentation processing of the cultural relic image of the cultural relic to be analyzed to obtain the image block set of the cultural relic image includes: determining the size of the image block according to the category to which the cultural relic to be analyzed belongs; performing image segmentation processing on the cultural relic image of the cultural relic to be analyzed according to the determined size of the image block to obtain the image block set of the cultural relic image. Among them, the size of the image block can be represented by the number of pixels included in the image block. According to experience and the resolution of the cultural relic image, the size of the image block corresponding to the category to which the cultural relic to be analyzed belongs can be pre-determined. By determining the size of the image block according to the category of the cultural relic, the size of the image block after the cultural relic image is segmented can be made more reasonable, neither too large nor too small, providing reliable data guarantee for subsequent processing.

[0035] In addition to determining the image block size based on the category to which the cultural relic belongs, other methods can also be used to determine the image block size. For example, by performing simple recognition and analysis on the cultural relic image, the richness of the information contained in the cultural relic image can be obtained, and the image block size can be determined according to the richness of the information. Or, image segmentation processing can also be performed through an AI model, and the AI model can be pre-trained to obtain a method for segmenting images. Based on this, the above-mentioned image segmentation processing of the cultural relic image of the cultural relic to be analyzed to obtain the image block set of the cultural relic image can include the following steps (1) and (2):

[0036] (1) Perform preliminary segmentation on the cultural relic image of the cultural relic to be analyzed through a first AI model to obtain multiple primary image blocks of a first size; wherein, the first AI model is pre-trained according to a first set of cultural relic image samples;

[0037] The cultural relic image in this embodiment can be a high-resolution image. The first AI model is used to perform preliminary understanding and segmentation on the cultural relic image, and the cultural relic image is divided into several primary image blocks of a first size.

[0038] The primary image blocks of the first size are related to the resolution of the cultural relic image. Generally, the higher the resolution of the cultural relic image, the larger the first size. In addition, the first size is also related to the richness of the information contained in the cultural relic image. Generally, the higher the richness of the information contained in the cultural relic image, the smaller the first size, that is, the more primary image blocks are obtained.

[0039] (2) Perform secondary segmentation on the multiple primary image blocks through the above-mentioned first AI model to obtain the image block set of the cultural relic image, wherein the image block size in the image block set is a second size, and the second size is smaller than the first size.

[0040] When performing secondary segmentation on the primary image blocks, the same secondary segmentation can be performed on each primary image block, or the first method of secondary segmentation can be performed on a part of the primary image blocks in a targeted manner, and the second method of secondary segmentation can be performed on another part of the primary image blocks, wherein the second size and / or the shape of the image blocks corresponding to the first method and the second method can be different.

[0041] The above-mentioned first set of cultural relic image samples can be composed of image samples pre-annotated with image segmentation methods.

[0042] The above method can make the segmentation of the cultural relic image more reasonable by using the first AI model for preliminary segmentation of preliminary image understanding and further detailed segmentation on this basis, which helps to reduce noise and errors in subsequent processing and is conducive to more accurately identifying the key feature areas on the surface of the cultural relic in the future.

[0043] In this embodiment, by dividing the cultural relic image into image blocks, each region of the cultural relic image can be segmented, which is beneficial to subsequently select the approximate region that is most likely to meet the analysis conditions, reduce the calculation amount, and then improve the calculation speed and the efficiency of feature extraction. If the computing power of the electronic device is relatively strong, the cultural relic image may not be segmented, and the image features of the cultural relic image can be directly extracted through the network model. However, this feature extraction method has a relatively large amount of computation and is prone to certain errors.

[0044] As a possible implementation manner, the shape of the image blocks in the above image block set is a square. The square image blocks can not only segment the cultural relic image relatively fully but also perform rotation operations during subsequent analysis and processing. In specific implementation, the image blocks can also adopt other shapes, such as rectangles or triangles, etc.

[0045] In order to more accurately evaluate the value of each image block in the cultural relic image, where this value can illustrate the role of the information contained in the image block in describing the cultural relic image, a set of standard evaluation indicators can be established in advance. The standard evaluation indicators can indicate that the score range of the significance evaluation indicators of the image blocks with specified type information is higher than the score range of the significance evaluation indicators of the image blocks without specified type information. The specified type information can be at least one of the following: texture features, frequency features, color change features, contrast features, edge features, etc.; if the information in the image block contains the specified type information, the significance evaluation indicator of the image block is determined from the first score range, and if the information in the image block does not contain any of the above specified type information, the significance evaluation indicator of the image block is determined from the second score range. Among them, the minimum value of the first score range is greater than the maximum value of the second score range.

[0046] To further simplify the operation of determining the significance evaluation index for the above-mentioned image blocks, the above standard evaluation index can also be determined by an AI model. The second AI model in this embodiment includes the standard evaluation index corresponding to the category to which the cultural relic to be analyzed belongs; this standard evaluation index is obtained by training the second AI model with a second set of cultural relic image samples; among them, the image blocks in the second set of cultural relic image samples are pre-annotated with significance evaluation index annotation values; based on this, determining the significance evaluation index of the image blocks in the image block set by using the pre-determined standard evaluation index of the category to which the cultural relic to be analyzed belongs can include: inputting the image blocks in the image block set into the second AI model to obtain the significance evaluation index of each image block output by the second AI model. During the process of training the second AI model, the historical cultural relic image block data can be screened, and the most representative image blocks can be selected for classification and marking. Here, the classification can include, from a large classification: calligraphy, seals, paintings, pottery, etc. Each item in this large classification can be further subdivided. For example, the calligraphy category can be subdivided into: subcategories such as the tails of left-falling strokes, right-falling strokes, rising strokes, and middle vertical strokes. Marking the image blocks can include marking the position area, classification, and the significance evaluation index annotation value corresponding to the image block in the entire cultural relic image. After classifying and marking the above-selected image blocks, a standard reference feature image block set can be formed, and this set can be used as the second set of cultural relic image samples to train the second AI model. The trained second AI model is used to evaluate the feature significance index of subsequent image blocks. This method of using an AI model to determine the significance evaluation index of image blocks can quickly analyze the features of image blocks and then quickly give the feature significance indexes of each image block, making the processing of a large amount of image data faster.

[0047] As a possible implementation manner, the significance evaluation index annotation value pre-annotated on the above image sample blocks is based on a pre-set feature dimension; among them, this feature dimension includes at least one of the following: texture feature, frequency feature, color change feature, contrast feature, edge feature. Through this annotation process, the second AI model trained with the image sample blocks can evaluate the image blocks more accurately, thus ensuring the accuracy of subsequent feature processing.

[0048] To further optimize the second AI model, the above method further includes: collecting the significance evaluation indicators of each image patch output by the second AI model; checking and adjusting the significance evaluation indicators of each image patch and adding them to the second cultural relic image sample set to update the second cultural relic image sample set; using the updated second cultural relic image sample set to optimize and train the second AI model. By checking and adjusting the significance evaluation indicators of the image patches output during the use of the second AI model, the inaccurate significance evaluation indicators output by the second AI model can be corrected more precisely. Then, the image patches after checking and adjusting the significance evaluation indicators are added to the second cultural relic image sample set, which can enrich the data in the second cultural relic image sample set by combining historical image patches (i.e., the original image sample patches in the second cultural relic image sample set) and the current image patches. On this basis, using the second cultural relic image sample set to iteratively optimize the standard evaluation indicators of the second AI model can improve the accuracy of the significance evaluation indicators output by the second AI model.

[0049] The above optimization training process of the second AI model can be carried out at regular intervals or when the number of collected image patches reaches a certain amount.

[0050] After obtaining the significance evaluation indicators of each image patch above, it can be seen which image patches are relatively important and which are relatively redundant according to the size of the significance evaluation indicators, and then the image patches can be screened. As a possible implementation manner, the method for screening out the feature patch set of the cultural relic image from the image patches in the image patch set according to the significance evaluation indicators of the image patches may include: (1) obtaining the significance threshold corresponding to the category of the cultural relic to be analyzed; this threshold can be determined in advance based on the category of the cultural relic and can be an empirical value, such as any value in 0.6 - 0.9. It can also be dynamically determined according to the significance evaluation indicators of each image patch corresponding to the cultural relic to be analyzed this time. For example, if there are 10 image patches, and the number of image patches with a significance evaluation indicator greater than 0.6 is 5, and the proportion of the number is 50% of the total number of image patches, then the significance threshold can be set to 0.6. (2) Screening out the target significance evaluation indicators greater than the significance threshold from the significance evaluation indicators of the image patches in the image patch set; (3) forming the feature patch set of the cultural relic image with the image patches corresponding to the target significance evaluation indicators. These selected feature patches contain the most critical information for understanding and analyzing the above cultural relic image. By screening the image patches based on the significance threshold, the image patches with high significance evaluation indicators can be used as feature patches, and the image patches with low significance evaluation indicators can be removed, reducing the number of image patches participating in subsequent feature processing, thereby reducing the computational amount of subsequent feature processing and ensuring the effect of feature processing.

[0051] The above-mentioned set of feature blocks can form a feature region of the cultural relic image. However, due to historical reasons, the cultural relic may be damaged to a certain extent. Therefore, AI technology can be used to accurately identify and complete the corresponding feature blocks to obtain a relatively complete feature region, and then mark the feature blocks in the feature region to distinguish them from other image blocks. On this basis, the determination of the feature information of the cultural relic image according to the set of feature blocks can include: extracting feature points from the feature blocks in the set of feature blocks through a third AI model; generating descriptors for each extracted feature point through the third AI model; wherein, the descriptor of the feature point is the feature information of the cultural relic image, and its specific form can be a digital vector of the key information of the cultural relic image, which can describe in detail the local image features around the feature block. The model structure and training process of the third AI model can be implemented with reference to the feature extraction model in related technologies, which will not be elaborated here. By extracting feature points from the feature blocks of the set of feature blocks and generating descriptors for each extracted feature point through the third AI model, the feature information in the feature region corresponding to the set of feature blocks can be quickly and accurately determined, providing reliable data for the identification and analysis of cultural relics.

[0052] To further understand the above method, taking the calligraphy and painting cultural relic image as an example, the processing method of the above-mentioned cultural relic image features includes the following steps:

[0053] I. Establishing a set of reference image blocks for historical calligraphy and painting images

[0054] This step mainly screens the image blocks corresponding to the key feature regions from the historical calligraphy and painting images.

[0055] First, segment the historical calligraphy and painting image, and identify and screen out the significant regions from the segmented image blocks. The significance evaluation index is used to determine which image blocks contain important feature information. The feature significance evaluation index can be determined based on various factors, such as texture complexity, color change, edge information, etc. These indexes help to identify the key regions in the image that are worthy of further analysis.

[0056] The above-mentioned historical calligraphy and painting images can come from the images on the surfaces of multiple different calligraphy and painting cultural relics, including various types of textures and features.

[0057] The image blocks in the above-mentioned selected significant feature regions are formed into a "set of reference image blocks". This set contains the most important image parts for the analysis of the surface of the calligraphy and painting cultural relic. For example, in a cultural relic image, certain regions may be selected as significant feature regions due to their unique texture patterns or damaged conditions.

[0058] II. Establishing a set of standard reference feature image blocks

[0059] Systematically classify and label the image patches in the above-mentioned set of reference images to obtain a set of standard reference feature image patches. By classifying and labeling the image patches in the set of reference images, a reference database (i.e., a set of standard reference feature image patches) containing a wide range of features related to calligraphy and painting cultural relics can be obtained. Each image patch in this database is assigned a clear label to describe its main features and attributes.

[0060] The set of standard reference feature image patches provides a comparison benchmark for the identification and evaluation of features in new images. By training the second AI model using the set of standard reference feature image patches, a set of significance evaluation criteria can be generated. When analyzing a new surface image of a cultural relic, its patches can be compared with the evaluation metrics generated using the patches in this standard set to evaluate the feature significance of each patch in the new image. This comparative analysis helps to determine which new image patches have unique or significant features.

[0061] III. Preliminary Understanding and Segmentation of Local Images of Calligraphy and Painting Cultural Relics

[0062] Taking Figure 2 the local image of the calligraphy and painting cultural relic shown as an example of the cultural relic image to be analyzed. The left side of this figure is the area where the ink is concentrated, and the lower right part is the rice paper area. The feature points that can reflect the characteristics of this image are concentrated in the intersection area of the ink and the rice paper. The pure black or pure white areas are difficult to provide effective information for the identification and analysis of cultural relics and thus do not belong to the feature points of this image.

[0063] In this step, artificial intelligence (AI, equivalent to the above-mentioned first AI model) first conducts a preliminary understanding of the obtained local image of the calligraphy and painting cultural relic. This step can use computer vision techniques, such as image recognition algorithms, to identify the key elements and features in the image. For example, the first AI model can identify cracks, texture changes, or other key visual features in the local image of the calligraphy and painting cultural relic.

[0064] After the preliminary understanding, the first AI model will perform a preliminary segmentation of the image based on the identified features. Specifically, the entire local image of the calligraphy and painting cultural relic is divided into larger blocks that contain key visual information, and each block centrally contains a certain specific texture or structure.

[0065] After the preliminary segmentation, the first AI model can further divide these larger blocks into smaller blocks. Refer to Figure 3 the schematic diagram of the segmented local image of the calligraphy and painting cultural relic shown. On the basis of the above Figure 2 the local image of the calligraphy and painting cultural relic is segmented into a set of 6×8 image patches for easier subsequent processing and analysis of specific regions of the image.

[0066] The size and shape of the sub - divisions can be flexibly adjusted according to the specific content and features of the image. For example, for areas with rich texture or complex details, smaller blocks may be needed to capture sufficient details; while for areas with relatively uniform texture, larger blocks can be used.

[0067] Through this segmentation and sub - division method, the subsequent feature extraction can be focused on the most important parts of the image, thereby improving the efficiency of the overall processing and the accuracy of the analysis. This is especially important for high - resolution images, because such images usually contain a large amount of data, and directly analyzing the entire image may be very time - consuming and prone to missing key details.

[0068] The above - mentioned segmentation method also has a certain degree of flexibility. The first AI model can dynamically adjust the size and shape of the blocks according to the specific content (such as the category it belongs to) and texture features of the cultural relic image. In this way, regardless of the complexity of the cultural relic image, the accuracy and effectiveness of the analysis can be guaranteed.

[0069] IV. Determination of Significance Evaluation Index

[0070] Generate descriptors for each image block of the local image of the painting and calligraphy cultural relic. A descriptor is a digital vector containing key information that can detail the local image features around the block. Use the above - mentioned second AI model to analyze these feature descriptors and evaluate the significance evaluation index of each image block. The significance evaluation index can be based on multiple dimensions such as the uniqueness, frequency, and contrast with the surrounding environment of the feature. Finally, the second AI model will output a significance evaluation index for each image block. This index reflects the importance and uniqueness level of the block in the entire image. See Figure 4 the schematic diagram of the evaluation index of the image blocks in the segmentation of the local image of the painting and calligraphy cultural relic shown. On the basis of the above Figure 3 , Figure 4 each image block in it is assigned a significance evaluation index, which reflects the importance and uniqueness of each image block in the overall image. This index is a value between 0 and 1. Among them, the significance evaluation index values of the image blocks corresponding to the ink - rice paper intersection area are relatively high, mostly greater than 0.9, while the significance evaluation index values of the image blocks in the pure - white and pure - black areas are relatively low, most of which are less than 0.2.

[0071] V. Automatic Determination of the Feature Block Set

[0072] After the second AI model outputs the saliency evaluation index for each image patch, in order to select the image patches with high value, a saliency threshold can be determined first. This threshold can be fixed or dynamically adjusted according to the specific content and requirements of the image. Then, all the image patches are traversed, and the image patches with saliency evaluation indexes higher than the saliency threshold are selected. These selected image patches contain the most critical information for understanding and analyzing the whole image. The selected image patches are combined into a set of feature patches, such as Figure 5 as shown in the schematic diagram of the local image feature area of the calligraphy and painting cultural relic, in which the saliency threshold is 0.9, and on the basis of Figure 4 the diagonal image patches are selected as feature patches, and the areas corresponding to these feature patches are the local image feature areas of the calligraphy and painting cultural relic, and this area will be the focus of subsequent in-depth analysis and processing.

[0073] VI. Analysis and Feature Point Extraction of Cultural Relic Images

[0074] In this step, AI technology can be applied to automatically identify and extract features from the feature patches in the set of feature patches, obtain feature points with significant structures, and then generate descriptors for each extracted feature point. These descriptors are digital vectors that can describe in detail the local image information around the feature points, and then subsequent analysis can be carried out. Continuing with the previous example, see as shown in Figure 6 the schematic diagram of the feature points of the local image of the calligraphy and painting cultural relic, where the small origin points are the feature points obtained by feature extraction on the basis of Figure 5 the above.

[0075] The above-mentioned second AI model can be optimized and trained periodically. During the use of the second AI model, historical image patch data are collected and sorted out, including the analysis results of past cultural relic images and the marked and classified feature areas, etc. Machine learning and artificial intelligence technologies are used to iteratively optimize the second AI model, including adjusting algorithm parameters, introducing new data features, etc., and training the second AI model with larger and better data samples.

[0076] Corresponding to the above method, an embodiment of the present disclosure also provides a device for processing cultural relic features. See Figure 7 , and this device includes the following modules:

[0077] A segmentation processing module 72, configured to perform image segmentation processing on the cultural relic image of the cultural relic to be analyzed, and obtain a set of image patches of the cultural relic image;

[0078] An evaluation index determination module 74, configured to determine the saliency evaluation index of the image patches in the set of image patches by applying the standard evaluation index of the category to which the cultural relic to be analyzed belongs determined in advance; wherein, the saliency evaluation index characterizes the richness of the information contained in the image patch in the cultural relic image;

[0079] A block screening module 76, configured to screen out a set of feature blocks of the cultural relic image from the set of image blocks according to the significance evaluation indexes of the image blocks in the set of image blocks of the cultural relic image;

[0080] A feature information determination module 78, configured to determine the feature information of the cultural relic image according to the set of feature blocks.

[0081] By performing image segmentation processing on the cultural relic image of the cultural relic to be analyzed, the above device can divide the cultural relic image into multiple smaller-sized image blocks, so that the information content and quantity included in each image block are different. On this basis, by applying the standard evaluation indexes of the category to which the cultural relic to be analyzed belongs that are determined in advance, the significance evaluation indexes of each image block are determined. The significance evaluation index is used to characterize the richness of the information included in the image block in the cultural relic image, so as to quantitatively measure each image block. Furthermore, according to the significance evaluation indexes of the image blocks, a set of feature blocks is screened out from the set of image blocks. The feature blocks in the set of feature blocks contain relatively rich information. Therefore, the feature blocks in the set of feature blocks can reflect the detailed information of the cultural relic image, and the set of feature blocks removes the image blocks in the set of image blocks that are not representative. While improving the accuracy of extracting the significant features on the surface of the cultural relic, it can also reduce the calculation amount of determining the feature information of the above cultural relic image and improve the efficiency of feature extraction, providing more in-depth and comprehensive cultural relic feature data for fields such as cultural relic protection, research, and identification.

[0082] As a possible implementation manner, the above segmentation processing module 72 is further configured to: determine the size of the image block according to the category to which the cultural relic to be analyzed belongs; perform image segmentation processing on the cultural relic image of the cultural relic to be analyzed according to the determined size of the image block to obtain a set of image blocks of the cultural relic image. By determining the size of the image block according to the category of the cultural relic, the size of the image block after the cultural relic image is segmented can be made more reasonable, neither too large nor too small, providing reliable data guarantee for subsequent processing.

[0083] As a possible implementation manner, the above segmentation processing module 72 is further configured to: perform preliminary segmentation on the cultural relic image of the cultural relic to be analyzed through a first AI model to obtain a plurality of primary image blocks with a first size; wherein, the first AI model is pre-trained according to a first set of cultural relic image samples;

[0084] The multiple primary image chunks are further segmented by the first AI model to obtain an image chunk set of the cultural relic image, wherein the image chunks in the image chunk set have a second size, and the second size is smaller than the first size. By using the first AI model for preliminary segmentation of preliminary image understanding and further detailed segmentation on this basis, the segmentation of the cultural relic image can be made more reasonable, which helps to reduce noise and errors in subsequent processing and is conducive to more accurately identifying the key feature regions on the surface of the cultural relics in the subsequent process.

[0085] As a possible implementation manner, the shape of the image chunks in the image chunk set is square. The square image chunks can not only make the segmentation of the cultural relic image relatively sufficient but also enable rotation operations in the subsequent analysis and processing.

[0086] As a possible implementation manner, the second AI model includes standard evaluation indicators corresponding to the category of the cultural relic to be analyzed; the standard evaluation indicators are obtained by training the second AI model with a second cultural relic image sample set; wherein, the image sample chunks in the second cultural relic image sample set are pre-labeled with significance evaluation indicator label values; correspondingly, the evaluation indicator determination module 74 is further configured to: input the image chunks in the image chunk set into the second AI model to obtain the significance evaluation indicators of each image chunk output by the second AI model. This method of using an AI model to determine the significance evaluation indicators of image chunks can quickly analyze the features of image chunks and then quickly give the feature significance indicators of each image chunk, which can make the processing of a large amount of image data faster.

[0087] As a possible implementation manner, the significance evaluation indicator label values pre-labeled on the image sample chunks are labeled based on a pre-set feature dimension; wherein, the feature dimension includes at least one of the following: texture feature, frequency feature, color change feature, contrast feature, edge feature. Through this labeling process, the second AI model trained by the image sample chunks can evaluate the image chunks more accurately, thereby ensuring the accuracy of subsequent feature processing.

[0088] As a possible implementation manner, the above-mentioned evaluation index determination module 74 is further configured to: collect the significance evaluation indexes of each image patch output by the second AI model; check and adjust the significance evaluation indexes of each image patch and then add them to the second cultural relic image sample set to update the second cultural relic image sample set; use the updated second cultural relic image sample set to optimize the training of the second AI model. By checking and adjusting the significance evaluation indexes of the image patches output during the use of the second AI model, the inaccurate significance evaluation indexes output by the second AI model can be corrected to be more accurate. Then, the image patches after checking and adjusting the significance evaluation indexes are added to the second cultural relic image sample set, which can enrich the data in the second cultural relic image sample set by combining historical image patches (i.e., the original image sample patches in the second cultural relic image sample set) and the current image patches. On this basis, using the second cultural relic image sample set to iteratively optimize the standard evaluation indexes of the second AI model can improve the accuracy of the significance evaluation indexes output by the second AI model.

[0089] As a possible implementation manner, the above-mentioned patch screening module 76 is further configured to: obtain the significance threshold corresponding to the category to which the cultural relic to be analyzed belongs; screen out the target significance evaluation indexes greater than the significance threshold from the significance evaluation indexes of the image patches in the image patch set; and form a feature patch set of the cultural relic image with the image patches corresponding to the target significance evaluation indexes. By screening image patches based on the significance threshold, the image patches with high significance evaluation indexes can be used as feature patches, and the image patches with low significance evaluation indexes can be removed, reducing the number of image patches participating in subsequent feature processing, thereby reducing the computational amount of subsequent feature processing and ensuring the effect of feature processing.

[0090] As a possible implementation manner, the above-mentioned feature information determination module 78 is further configured to: extract feature points from the feature patches in the feature patch set through a third AI model; and generate descriptors for each extracted feature point through the third AI model; wherein, the descriptor of the feature point is the feature information of the cultural relic image. By extracting feature points from the feature patches in the feature patch set and generating descriptors for each extracted feature point through the third AI model, the feature information in the feature region corresponding to the feature patch set can be quickly and accurately determined, providing reliable data for cultural relic identification and analysis.

[0091] The implementation principle and the technical effects generated by a cultural relic feature processing device provided by an embodiment of the present disclosure are the same as those of the foregoing method embodiment. For a brief description, for the parts not mentioned in the embodiment of the cultural relic feature processing device, reference may be made to the corresponding content in the foregoing cultural relic feature processing embodiment.

[0092] As used herein, the term "and / or" is merely a description of the associated relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent three situations: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the term "at least one" as used herein means any one of a plurality or any combination of at least two of a plurality. For example, including at least one of A, B, and C may mean including any one or more elements selected from the set consisting of A, B, and C.

[0093] Embodiments of the present disclosure also provide an electronic device, such as Figure 8 shown, which is a schematic structural diagram of the electronic device. Among them, the electronic device includes a processor 81 and a memory 80. The memory 80 stores computer-executable instructions that can be executed by the processor 81, and the processor 81 executes the computer-executable instructions to implement the above-mentioned processing method for cultural relic features.

[0094] In Figure 8 the illustrated embodiment, the electronic device further includes a bus 82 and a communication interface 83. Among them, the processor 81, the communication interface 83, and the memory 80 are connected through the bus 82.

[0095] Among them, the memory 80 may include a high-speed random access memory (RAM, Random Access Memory), and may also include a non-volatile memory, such as at least one disk memory. Through at least one communication interface 83 (which may be wired or wireless), a communication connection is established between the system network element and at least one other network element, and the Internet, wide area network, local area network, metropolitan area network, etc. can be used. The bus 82 may be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus 82 can be divided into an address bus, a data bus, a control bus, etc. For the sake of representation, Figure 8 only a bidirectional arrow is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0096] The processor 81 may be an integrated circuit chip with signal processing capabilities. In the implementation process, the steps of the above method can be completed by the integrated logic circuit of the hardware in the processor 81 or the instructions in the form of software. The above-mentioned processor 81 may be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it may also be a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field-programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present disclosure can be directly embodied as being executed and completed by the hardware decoding processor, or executed and completed by a combination of the hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory, and the processor 81 reads the information in the memory and combines its hardware to complete the steps of the method for processing the cultural relic features in the foregoing embodiments.

[0097] The embodiments of the present disclosure also provide a computer-readable storage medium storing computer-executable instructions, which, when called and executed by a processor, cause the processor to implement the above method for processing cultural relic features. For specific implementation, reference may be made to the foregoing method embodiments, and details are not described herein again.

[0098] The computer program product of the method, device, and electronic device for processing cultural relic features provided by the embodiments of the present disclosure includes a computer-readable storage medium storing program code, and the instructions included in the program code can be used to execute the method described in the foregoing method embodiments. For specific implementation, reference may be made to the method embodiments, and details are not described herein again.

[0099] Unless otherwise specifically stated, the relative steps, numerical expressions, and values of the components and steps set forth in these embodiments do not limit the scope of the present disclosure.

[0100] If the above-mentioned functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium executable by a processor. Based on this understanding, the technical solution of the present disclosure, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present disclosure. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0101] In the description of the present disclosure, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present disclosure and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation to the present disclosure. In addition, the terms "first", "second", "third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0102] Finally, it should be noted that the above-mentioned embodiments are only specific embodiments of the present disclosure, used to illustrate the technical solutions of the present disclosure, rather than limiting them. The protection scope of the present disclosure is not limited thereto. Although the present disclosure has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: any person skilled in the art within the technical scope disclosed by the present disclosure can still modify the technical solutions recorded in the foregoing embodiments or can easily think of changes, or perform equivalent replacements on some of the technical features; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present disclosure, and should all be covered within the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be determined by the protection scope of the claims.

Claims

1. A method for processing cultural relic features, characterized in that The method includes: Performing image segmentation processing on the cultural relic image of the cultural relic to be analyzed to obtain an image block set of the cultural relic image; Applying a standard evaluation index of the category to which the cultural relic to be analyzed belongs, which is determined in advance, to determine a significance evaluation index of the image blocks in the image block set; wherein, the significance evaluation index characterizes the richness of the information contained in the image block in the cultural relic image; Screening out a feature block set of the cultural relic image from the image block set according to the significance evaluation index of the image blocks in the image block set; Determining the feature information of the cultural relic image according to the feature block set.

2. The method according to claim 1, wherein Performing image segmentation processing on the cultural relic image of the cultural relic to be analyzed to obtain an image block set of the cultural relic image includes: Determining the size of the image blocks according to the category to which the cultural relic to be analyzed belongs; Performing image segmentation processing on the cultural relic image of the cultural relic to be analyzed according to the determined size of the image blocks to obtain an image block set of the cultural relic image.

3. The method according to claim 1, characterized in that, Performing image segmentation processing on the cultural relic image of the cultural relic to be analyzed to obtain an image block set of the cultural relic image includes: Performing preliminary segmentation on the cultural relic image of the cultural relic to be analyzed through a first AI model to obtain a plurality of primary image blocks of a first size; wherein, the first AI model is trained in advance according to a first cultural relic image sample set; Performing re-segmentation on the plurality of primary image blocks through the first AI model to obtain an image block set of the cultural relic image, wherein the size of the image blocks in the image block set is a second size, and the second size is smaller than the first size.

4. The method according to any one of claims 1 to 3, characterized in that The shape of the image blocks in the image block set is square.

5. The method according to claim 1, wherein The second AI model contains a standard evaluation index corresponding to the category to which the cultural relic to be analyzed belongs; the standard evaluation index is obtained by training the second AI model with a second cultural relic image sample set; wherein, the image sample blocks in the second cultural relic image sample set are pre-labeled with significance evaluation index label values; The applying a standard evaluation index of the category to which the cultural relic to be analyzed belongs, which is determined in advance, to determine a significance evaluation index of the image blocks in the image block set includes: Inputting the image blocks in the image block set into the second AI model to obtain the significance evaluation index of each image block output by the second AI model.

6. The method according to claim 5, characterized in that, The significance evaluation index label values pre-labeled on the image sample blocks are labeled based on a pre-set feature dimension; wherein, the feature dimension includes at least one of the following: texture feature, frequency feature, color change feature, contrast feature, edge feature.

7. The method according to claim 5, wherein The method further includes: Collecting the significance evaluation index of each image block output by the second AI model; Checking and adjusting the significance evaluation index of each image block and adding it to the second cultural relic image sample set to update the second cultural relic image sample set; Optimizing and training the second AI model using the updated second cultural relic image sample set.

8. The method according to claim 1, wherein Based on the significance evaluation index of the image patches in the image patch set, screening out the feature patch set of the cultural relic image from the image patch set includes: Obtaining the significance threshold corresponding to the category to which the cultural relic to be analyzed belongs; Screening out the target significance evaluation index greater than the significance threshold from the significance evaluation indexes of the image patches in the image patch set; Forming the feature patch set of the cultural relic image with the image patches corresponding to the target significance evaluation index.

9. The method according to claim 1, characterized in that Determining the feature information of the cultural relic image according to the feature patch set includes: Extracting feature points from the feature patches in the feature patch set through a third AI model; Generating descriptors for each of the extracted feature points through the third AI model; wherein, the descriptors of the feature points are the feature information of the cultural relic image.

10. A processing device for cultural relic features, characterized in that, The device includes: A segmentation processing module, configured to perform image segmentation processing on the cultural relic image of the cultural relic to be analyzed to obtain the image patch set of the cultural relic image; An evaluation index determination module, configured to determine the significance evaluation index of the image patches in the image patch set by applying the standard evaluation index of the category to which the cultural relic to be analyzed belongs; wherein, the significance evaluation index characterizes the richness of the information contained in the image patch in the cultural relic image; A patch screening module, configured to screen out the feature patch set of the cultural relic image from the image patch set according to the significance evaluation index of the image patches in the image patch set; A feature information determination module, configured to determine the feature information of the cultural relic image according to the feature patch set.

11. An electronic device, characterized in that, It includes a processor and a memory, the memory stores computer-executable instructions that can be executed by the processor, and the processor executes the computer-executable instructions to implement the method according to any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are called and executed by a processor, the computer-executable instructions cause the processor to implement the method according to any one of claims 1 to 9.

Images