Garment insurance loss assessment method and device based on vision, equipment and medium

Abstract

The invention relates to a vision-based garment insurance loss assessment method and device, equipment and a medium. According to the method, the garment type corresponding to the target garment is determined, the current visual feature matrix and the current physical feature matrix are fused, the joint feature vector and the original feature vector corresponding to the target garment are compared, the comparison result is quantized and mapped with the loss level, and the loss level of the target garment is calculated. And determining a loss assessment result corresponding to the target garment according to the predicted loss level and the insurance policy information corresponding to the target garment. The method can be applied to a financial insurance business scene, the image, fabric and tensile strength information of the target garment is acquired through the image acquisition device, the spectrometer device and the tension sensor, the comparison result is quantized and mapped into the loss level, and finally the loss assessment result of the target garment is determined in combination with insurance policy information. Therefore, the clothing image is comprehensively subjected to multi-dimensional analysis to determine the loss level so as to carry out loss assessment on the clothing.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of artificial intelligence and the field of financial services, in particular to a clothing insurance loss assessment method and device based on vision, equipment and medium. BACKGROUND

[0002] In the field of clothing insurance, accurate loss assessment is a key link to protect the rights and interests of both parties. The traditional clothing insurance loss assessment method mainly relies on manual inspection and experience judgment. However, with the rapid development of the clothing industry and the continuous expansion of insurance business, this traditional method gradually exposes many drawbacks.

[0003] Traditional loss assessment relies heavily on manual review. When facing special targets such as high-end custom-made clothing, it is easy to cause misjudgment due to experience differences, and it is difficult to quantify subtle losses such as loose embroidery and pilling of fabrics. Although image editing and generation technology can achieve appearance replacement, it lacks the ability to quantify key parameters of insurance business (such as material loss degree and repair cost), and there are problems such as background information interference and inaccurate feature migration, which cannot directly adapt to complex insurance assessment scenarios. For example, in the field of property insurance, the assessment ability of clothing tearing, burning and internal structure damage is weak, and the lack of deep integration with insurance loss assessment rules makes it difficult to directly convert technical results into claims settlement decisions, and the overall efficiency cannot meet the fine needs of high-end item insurance.

[0004] Therefore, how to comprehensively analyze the clothing image in multiple dimensions to determine the loss level and assess the clothing has become a problem to be solved. SUMMARY

[0005] Therefore, the embodiments of the present application provide a clothing insurance loss assessment method and device based on vision, equipment and medium to solve the problem of how to comprehensively analyze the clothing image in multiple dimensions to determine the loss level and assess the clothing.

[0006] In a first aspect, the embodiments of the present application provide a clothing insurance loss assessment method based on vision, comprising: Based on the image acquisition device, the current image information of the target clothing is acquired, based on the spectrometer device, the current fabric information of the target clothing is acquired, based on the tensile force sensor, the current tensile strength information of the target clothing is acquired, and according to the current image information, the clothing type corresponding to the target clothing is determined; According to the current image information, the current visual feature vector representing the stitches, colors and textures of the target clothing is constructed, and according to the current fabric information and the current tensile strength information, the current physical feature vector representing the fiber density and fabric elasticity is constructed; Based on the clothing type, a weighting factor is determined. Based on the weighting factor, the current visual feature matrix and the current physical feature matrix are fused to obtain a joint feature vector. The joint feature vector is then compared with the original feature vector corresponding to the target clothing to determine the comparison result. The comparison results are quantified and mapped to the loss level to obtain the predicted loss level. Based on the predicted loss level and the corresponding policy information of the target garment, the loss assessment result for the target garment is determined.

[0007] Secondly, an embodiment of this application provides a vision-based clothing insurance loss assessment device, comprising: The clothing information acquisition module is used to acquire image information of the target clothing based on an image acquisition device, acquire fabric information of the target clothing based on a spectrometer, acquire tensile strength information of the target clothing based on a tensile sensor, and determine the clothing type corresponding to the target clothing based on the image information. The feature vector analysis module is used to construct visual feature vectors representing the stitches, colors, and textures of the target garment based on the image information, and to construct physical feature vectors representing fiber density and fabric elasticity based on the fabric information and tensile strength information. The feature comparison module is used to determine a weighting factor based on the clothing type, fuse the visual feature matrix and the physical feature matrix based on the weighting factor to obtain a joint feature vector, and compare the joint feature vector with the original feature vector corresponding to the target clothing to determine the comparison result. The insurance loss assessment module is used to quantify the comparison results and map them to the loss level to obtain the predicted loss level. Based on the predicted loss level and the corresponding policy information of the target garment, the loss assessment result for the target garment is determined.

[0008] Thirdly, embodiments of this application provide a computer device, the computer device including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the vision-based clothing insurance loss assessment method as described in the first aspect.

[0009] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the vision-based clothing insurance loss assessment method as described in the first aspect.

[0010] The beneficial effects of the embodiments in this application compared with the prior art are: In this application, current image information of the target garment is acquired using an image acquisition device, current fabric information of the target garment is acquired using a spectrometer, and current tensile strength information of the target garment is acquired using a tensile sensor. Based on the current image information, the garment type of the corresponding target garment is determined. Based on the current image information, a current visual feature vector representing the stitching, color, and texture of the target garment is constructed. Based on the current fabric information and current tensile strength information, a current physical feature vector representing fiber density and fabric elasticity is constructed. Based on the garment type, a weighting factor is determined. Based on the weighting factor, the current visual feature matrix and the current physical feature matrix are fused to obtain a joint feature vector. The joint feature vector is compared with the original feature vector corresponding to the target garment to determine the comparison result. The comparison result is quantified and mapped with the loss level to obtain the predicted loss level. Based on the predicted loss level and the corresponding policy information of the target garment, the loss assessment result of the corresponding target garment is determined. This application can be applied to insurance business scenarios. It uses image acquisition equipment, a spectrometer, and a tensile sensor to collect images, fabric information, and tensile strength information of the target garment. Then, the garment type is determined, and visual and physical feature vectors are constructed. Based on the garment type, weighting factors are determined to fuse the two types of features to obtain a joint feature vector. This joint feature vector is compared with the original feature vector, the comparison result is quantified, and mapped to a loss level. Finally, combined with policy information, the loss assessment result for the target garment is determined. This allows for comprehensive, multi-dimensional analysis of garment images to determine the loss level and assess the garment's damage. BRIEF DESCRIPTION OF DRAWINGS

[0011] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of an application environment for a vision-based clothing insurance loss assessment method provided in Embodiment 1 of this application; Figure 2 This is a flowchart illustrating a vision-based clothing insurance loss assessment method provided in Embodiment 2 of this application; Figure 3 This is a flowchart illustrating a vision-based clothing insurance loss assessment method provided in Embodiment 3 of this application; Figure 4 This is a schematic flowchart of a vision-based clothing insurance loss assessment method provided in Embodiment 4 of this application; Figure 5 This is a flowchart illustrating a vision-based clothing insurance loss assessment method provided in Embodiment 5 of this application; Figure 6 This is a schematic diagram of a vision-based clothing insurance loss assessment device provided in Embodiment Six of this application; Figure 7 This is a schematic diagram of the structure of a computer device provided in Embodiment 7 of this application. DETAILED DESCRIPTION

[0013] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0014] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0015] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0016] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0017] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0018] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0019] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0020] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0021] It should be understood that the sequence number of each step in the following embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0022] To illustrate the technical solution of this application, specific embodiments are described below.

[0023] The first embodiment of this application provides a vision-based method for assessing losses in clothing insurance, which can be applied to applications such as... Figure 1 In this application environment, the client and server communicate with each other. Users can provide conditions, requirements, and operation instructions for vision-based clothing insurance loss assessment through the client. The server is used to provide control instructions for the vision-based clothing insurance loss assessment method based on the relevant content sent by the client.

[0024] The aforementioned vision-based clothing insurance loss assessment method can be applied to, but is not limited to, financial systems, such as banking and securities systems. System development can be achieved using a Software Development Kit (SDK), specifically developing modules such as application authentication, intelligent routing, and high availability management. Client devices include, but are not limited to, handheld computers, desktop computers, laptops, ultra-mobile personal computers (UMPCs), netbooks, cloud terminal devices, and personal digital assistants (PDAs). Server devices can be independent servers or cloud servers providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0025] See Figure 2 This is a flowchart illustrating a vision-based clothing insurance loss assessment method provided in Embodiment 2 of this application. The aforementioned vision-based clothing insurance loss assessment method can be applied to... Figure 1 The server-side component.

[0026] like Figure 2 As shown, this vision-based clothing insurance loss assessment method may include the following steps: Step S201: Acquire current image information of the target garment based on the image acquisition device, acquire current fabric information of the target garment based on the spectrometer device, acquire current tensile strength information of the target garment based on the tensile sensor, and determine the garment type corresponding to the target garment based on the current image information.

[0027] Image acquisition equipment refers to various types of cameras, such as ordinary digital SLR cameras and mobile phone cameras. In practical applications, devices with appropriate resolution and shooting performance are selected according to specific needs to ensure that clear and accurate images of clothing are captured.

[0028] Image information contains a wealth of visual information, such as the overall style of the clothing (whether it's a dress, suit, or casual wear), color distribution, pattern, and whether there are any damages or stains on the surface. This information is crucial for judging the appearance of the clothing and subsequently determining its type. For example, images can clearly show features such as the neckline design and hem shape, thus aiding in the identification of the clothing type. A spectrometer is an instrument that can analyze the spectral properties of a substance. In the context of clothing damage assessment, it can obtain relevant information about the fabric by detecting its absorption and reflection of light at different wavelengths.

[0029] Fabric information includes the fabric's composition (such as whether it's a single component like pure cotton, polyester, or wool, or a blend of multiple components) and fiber structure. Different fabric compositions and structures affect the quality, durability, and value of clothing. For example, pure cotton fabrics are highly absorbent but relatively prone to deformation, while wool-containing fabrics are very warm but may require special care. Understanding fabric information helps in more accurately assessing the degree of damage and the assessed value of clothing.

[0030] A tension sensor is a device that converts tensile force signals into electrical signals. When collecting information on the tensile strength of clothing, a certain tensile force is applied to a part of the clothing, and the tension sensor measures and records the changes in tensile force in real time during the stress process.

[0031] Tensile strength reflects a garment fabric's ability to resist tensile failure. As garments are used and worn, their tensile strength may change. For example, after multiple washes or prolonged wear, the fabric fibers may be damaged, leading to a decrease in tensile strength. By collecting current tensile strength information, we can understand the physical performance state of the garment, providing an important basis for judging its degree of wear and tear. Computer vision technology and pattern recognition algorithms are used to process and analyze the collected image information. Deep learning-based image classification models, trained on a large number of images of different types of garments, can learn the characteristic patterns of various garment types.

[0032] Different types of clothing vary in materials, manufacturing processes, and usage scenarios, and their wear and tear conditions and damage assessment standards may also differ. Accurately identifying the clothing type helps in subsequent steps to set appropriate weighting factors for different types of clothing, more accurately assessing the degree of wear and tear and determining the damage assessment results. For example, the focus of damage assessment for suits and sweatpants may differ; suits emphasize the integrity of appearance and fit, while sweatpants may focus more on the abrasion resistance and tensile strength of the fabric.

[0033] For example, in a property insurance scenario, the insurance company arranges for claims adjusters to carry high-definition cameras to a designated location to photograph the damaged suit from multiple angles, capturing clear images of the suit's overall appearance and close-ups of the scratched areas. The claims adjusters then use a portable spectrometer to analyze the suit fabric, obtaining information such as its composition and fiber structure, confirming that the fabric is a top-grade wool blend. Using a small tensile sensor, the claims adjusters test the tensile strength of the scratched area and surrounding region, obtaining current tensile strength information and finding that the tensile strength at the scratched area has decreased. Based on the collected image information, image recognition software automatically identifies the garment as a men's suit.

[0034] Step S202: Based on the current image information, construct a current visual feature vector representing the stitches, color, and texture of the target garment; based on the current fabric information and the current tensile strength information, construct a current physical feature vector representing the fiber density and fabric elasticity.

[0035] Among these steps, stitch feature extraction involves using image processing algorithms (such as edge detection and morphological operations) to identify the position and shape of stitches from the current image information. The stitch features are then quantified, for example, by using the average or variance of the stitch spacing as feature values. A large variance in the stitch spacing indicates uneven stitching, which may suggest problems in the garment's manufacturing process or during use.

[0036] Color feature extraction can employ color space conversion methods, transforming images from the common RGB color space to more suitable color analysis spaces such as HSV (hue, saturation, brightness). This involves statistically analyzing the distribution of different colors within the image.

[0037] Texture feature extraction can utilize texture analysis algorithms, such as gray-level co-occurrence matrix (GLCM) and local binary pattern (LBP), to describe the characteristics of clothing textures. GLCM reflects the spatial distribution of gray levels in an image, thereby extracting information such as texture roughness and contrast.

[0038] The extracted texture features are quantized to form feature values. For example, the magnitude of texture roughness can be used as a feature value, and changes in roughness may be related to the degree of wear and tear on clothing.

[0039] The feature values ​​of stitching, color, and texture are combined to form a multi-dimensional vector that represents the current visual characteristics of the target garment. Each element in the vector corresponds to a specific feature value, and this vector comprehensively describes the visual appearance features of the garment.

[0040] Physical feature vectors reflect the state of the target garment from the perspective of physical performance. Fiber density and fabric elasticity are important physical indicators that affect the quality and durability of garments. Constructing relevant feature vectors helps to assess the physical wear and tear of garments.

[0041] Fiber density characteristics are obtained by analyzing the fabric composition and structure based on current fabric information collected by a spectrometer, and then inferring the fiber arrangement and density. Different fabric compositions and structures will affect the measurement and calculation of fiber density.

[0042] Fiber density can be expressed as the number or weight of fibers per unit area. Variations in fiber density may be related to fabric wear, shrinkage, and other factors.

[0043] Fabric elasticity characteristics can be acquired using tensile strength information collected by tensile sensors to analyze the fabric's deformation and recovery ability under stress. For example, the elongation during stretching and the recovery rate after unloading can be recorded. Elasticity-related characteristics can be quantified using metrics such as elastic modulus and elongation at break. Elastic modulus reflects the fabric's resistance to elastic deformation, while elongation at break represents the maximum proportion of stretch the fabric can achieve before breaking.

[0044] The characteristic values ​​of fiber density and fabric elasticity are combined into a vector, which represents the current physical characteristic vector of the target garment. This current physical characteristic vector allows us to understand the physical performance state of the garment, providing a basis for subsequent damage assessment.

[0045] For example, in an insurance claims settlement scenario, insurance company technicians analyze collected images to extract features such as the evenness of stitching, color consistency, and texture integrity of the suit, constructing a current visual feature vector. They discover that some stitches are broken at the scratched area, the color changes slightly at the scratch, and the texture is also damaged. Combining fabric information and tensile strength information, they construct a current physical feature vector representing fiber density and fabric elasticity. Analysis shows that the fiber density at the scratched area is reduced, and the fabric elasticity is also less than in undamaged areas.

[0046] Step S203: Determine a weighting factor based on the clothing type; fuse the current visual feature matrix and the current physical feature matrix based on the weighting factor to obtain a joint feature vector; compare the joint feature vector with the original feature vector corresponding to the target clothing to determine the comparison result.

[0047] Different types of clothing have varying degrees of importance in terms of visual and physical characteristics, so the weighting factors must be determined based on the clothing type identified in step S201. For example, for evening gowns, which emphasize appearance design, visual features are more crucial for assessing their overall value and wear and tear, so the weight of visual features will be higher. On the other hand, for workwear, which emphasizes practicality and durability, physical features (such as fabric strength and abrasion resistance) have a greater weight in the assessment, so the weight of physical features will be relatively higher.

[0048] Weighting factors can be determined through expert experience, historical data statistical analysis, or machine learning algorithms. Experts, with their extensive industry knowledge and experience, can provide reasonable weight allocations based on the characteristics of different clothing types. Historical data statistical analysis involves studying a large number of damage assessment cases of similar clothing to identify the degree of influence of visual and physical features on the final damage assessment, thereby determining the weights. Machine learning algorithms can be trained on a large amount of labeled data on different types of clothing to automatically learn the appropriate weight allocations for different clothing types.

[0049] In step S202, the current visual feature vector and the current physical feature vector are constructed. To facilitate the fusion operation, these vectors are usually extended into matrix form (in simple cases, vectors can also be regarded as special matrices). The current visual feature matrix contains visual feature information such as stitching, color, and texture, while the current physical feature matrix contains physical feature information such as fiber density and fabric elasticity. Based on the determined weighting factors, the current visual feature matrix and the current physical feature matrix are weighted and combined. Assuming the visual feature matrix is ​​V, the physical feature matrix is ​​P, the weighting factor for the visual features is wv, and the weighting factor for the physical features is wp, then the joint feature vector F can be calculated using the following formula: F = wV + wpp. Through this weighted fusion method, both visual and physical features are comprehensively considered, enabling the joint feature vector to more comprehensively reflect the current state of the target garment. The original feature vector is obtained by fusing image information, fabric information, and tensile strength information collected in the same way as in the current step when the target garment is in its brand-new state. This process constructs a visual feature matrix and a physical feature matrix. It represents the initial, ideal state of the garment. The cosine similarity value ranges from -1 to 1. A value closer to 1 indicates greater similarity between the two vectors, meaning the current state of the target garment is closer to its initial state, and the degree of damage is lower. A value closer to -1 indicates greater difference between the two vectors, and a higher degree of damage to the garment. Euclidean distance can also be used; a smaller Euclidean distance indicates closer vectors and less damage to the garment, while a larger distance indicates greater damage. The result obtained through comparison is a specific numerical value that intuitively reflects the degree of change in the target garment's current state relative to its initial state. Subsequent steps will quantify the degree of damage to the garment based on this comparison result and map it to a damage level to determine the final damage assessment result.

[0050] For example, in an insurance claims settlement scenario, the weight of the suit's visual features is set to 0.6, and the weight of its physical features is set to 0.4. Based on these weighting factors, the current visual feature matrix and the current physical feature matrix are fused to obtain a joint feature vector. The insurance company has already collected and saved the original feature vector of the suit when the customer applies for insurance. Comparing the joint feature vector with the original feature vector reveals significant differences, determining the specific comparison results, such as a 30% difference in visual features and a 20% difference in physical features.

[0051] Step S204: Quantify the comparison result and map it to the loss level to obtain the predicted loss level. Based on the predicted loss level and the corresponding policy information of the target garment, determine the loss assessment result for the target garment.

[0052] If the comparison result is a continuous numerical range, such as a cosine similarity value between -1 and 1, it can be mapped to a specific quantization interval, such as the integer interval between 0 and 100, through a linear transformation. Based on the distribution of the comparison results, they are divided into different intervals, each corresponding to a specific quantization value. For example, when the Euclidean distance is between 0 and 10, the quantization value is 10; when it is between 11 and 20, the quantization value is 20, and so on.

[0053] Different damage levels are pre-defined based on the potential wear and tear on the clothing, such as minor damage, moderate damage, and severe damage. Each damage level corresponds to a range of quantified values. For example, a quantified value between 0 and 20 corresponds to minor damage, between 21 and 50 corresponds to moderate damage, and between 51 and 100 corresponds to severe damage.

[0054] The quantified comparison results are matched with a pre-defined loss level range to determine the predicted loss level of the target garment. For example, if the quantification value is 30, the predicted loss level of the garment is moderate loss.

[0055] Policy information refers to the terms and parameters stipulated in the insurance contract signed between the insurance company and the policyholder, including the sum insured, compensation ratio, deductible, etc. Different clothing insurance policies may have different provisions. This information is an important basis for determining the loss assessment result.

[0056] Based on the predicted level of damage, find the corresponding compensation ratio in the policy. Assuming the insured amount is A, and the compensation ratio corresponding to the predicted level of damage is r, the assessed loss D can be calculated using the formula D = A × r. For example, if the insured amount is 1000 yuan, and the compensation ratio for minor damage is 20%, then the assessed loss would be 1000 x 0.2 = 200 yuan.

[0057] Some insurance policies stipulate a deductible, meaning that the insurance company will not be liable for compensation if the loss amount is below a certain value. Assuming the deductible is B, if the calculated compensation amount D is less than B, the loss assessment result is 0; if D is greater than B, the actual compensation amount is DB.

[0058] For example, in an insurance claims scenario, the quantification and mapping of damage levels involves quantifying the comparison results mentioned above to arrive at a comprehensive quantified value. Based on the insurance company's pre-set damage level standards, this quantified value is mapped to the corresponding damage level, determining that the suit's damage level is moderate. Checking the customer's policy information, it is found that the compensation ratio for moderate damage is 40%. Considering the suit's insured value of 50,000 yuan, the final assessment result is a compensation of 20,000 yuan to the customer.

[0059] In this application, current image information of the target garment is acquired using an image acquisition device, current fabric information of the target garment is acquired using a spectrometer, and current tensile strength information of the target garment is acquired using a tensile sensor. Based on the current image information, the garment type of the corresponding target garment is determined. Based on the current image information, a current visual feature vector representing the stitching, color, and texture of the target garment is constructed. Based on the current fabric information and current tensile strength information, a current physical feature vector representing fiber density and fabric elasticity is constructed. Based on the garment type, a weighting factor is determined. Based on the weighting factor, the current visual feature matrix and the current physical feature matrix are fused to obtain a joint feature vector. The joint feature vector is compared with the original feature vector corresponding to the target garment to determine the comparison result. The comparison result is quantified and mapped with the loss level to obtain the predicted loss level. Based on the predicted loss level and the corresponding policy information of the target garment, the loss assessment result of the corresponding target garment is determined. This application can be applied to insurance business scenarios. It uses image acquisition equipment, a spectrometer, and a tensile sensor to collect images, fabric information, and tensile strength information of the target garment. Then, the garment type is determined, and visual and physical feature vectors are constructed. Based on the garment type, weighting factors are determined to fuse the two types of features to obtain a joint feature vector. This joint feature vector is compared with the original feature vector, the comparison result is quantified, and mapped to a loss level. Finally, combined with policy information, the loss assessment result for the target garment is determined. This allows for comprehensive, multi-dimensional analysis of garment images to determine the loss level and assess the garment's damage.

[0060] See Figure 3 This is a flowchart illustrating a vision-based clothing insurance loss assessment method provided in Embodiment 3 of this application. Figure 3 As shown, before comparing the joint feature vector with the original feature vector corresponding to the target garment to determine the comparison result in step S203 above, the following steps may also be included: Step S301: Based on the current image information, determine the target contour features corresponding to at least one target feature region in the target clothing.

[0061] Step S302: Obtain the original image information of the target garment, determine the original contour features of the corresponding target feature region from the original image information, and determine the mapping matrix representing the deformation based on the target contour features and the original contour features.

[0062] Step S303: Using the mapping matrix, the coordinates of each feature point in the original image information of the target garment are aligned to obtain the aligned feature point coordinates.

[0063] Step S304: Based on the aligned feature point coordinates, the sample feature vector of the target garment is corrected to obtain the original feature vector corresponding to the target garment.

[0064] Optionally, the following steps may be included before step S304: Obtain the original fabric information and original tensile strength information of the target garment; Based on the original image information, an original visual feature vector representing the stitches, color, and texture of the target garment is constructed; based on the original fabric information and the original tensile strength information, an original physical feature vector representing the fiber density and fabric elasticity is constructed. Based on the weighting factor, the original visual feature matrix and the original physical feature matrix are fused to obtain the sample feature vector.

[0065] Optionally, step S303 may include the following steps: Using the moving least squares method, based on the mapping matrix, the coordinates of each feature point in the original image information of the target garment are mapped and aligned to obtain the aligned feature point coordinates.

[0066] The target garment will deform during use. To measure this deformation, it is necessary to first identify some key feature regions and their contour features within the garment. These feature regions are typically representative, easily identifiable, and significant parts of the garment in deformation analysis, such as the collar, cuffs, and pockets. Based on the current image information, image processing techniques (such as edge detection and contour extraction algorithms) are used to identify at least one target feature region in the garment and extract the contour features of these regions. For example, the Canny edge detection algorithm can be used to find the edges of the target feature regions, and then a contour tracking algorithm can be used to obtain the shape, size, position, and other feature information of their contours.

[0067] By comparing the current target contour features of the garment with the original contour features, the deformation of the garment is determined, and this deformation is represented by a mapping matrix. This mapping matrix describes the transformation relationship of feature points from the original state to the current state.

[0068] The original contour features can be obtained by acquiring the original image information of the target garment and using a method similar to step S301 to determine the original contour features of the corresponding target feature region from the original image information.

[0069] Calculating the mapping matrix allows for a comparative analysis of the target contour features and the original contour features. Using a suitable algorithm (such as affine transformation or perspective transformation), a mapping matrix describing the deformation relationship between the two is calculated. This matrix maps feature points in the original image to their current deformed positions. Using the mapping matrix obtained in step S302, the coordinates of each feature point in the original image of the target clothing are adjusted to align with the current deformed state, thereby eliminating the influence of deformation on the feature point positions and preparing for subsequent correction of the original feature vector.

[0070] Moving least squares is an interpolation and fitting method that aligns feature points in an original image using a mapping matrix. This method smoothly transforms feature points locally, making the aligned points more consistent with actual deformations. Based on moving least squares, the mapping matrix is ​​applied to the coordinates of each feature point in the original image information to obtain the aligned feature point coordinates.

[0071] Based on the aligned feature point coordinates, the sample feature vector of the target garment is corrected to obtain an original feature vector that more accurately reflects the original state of the garment. This is because the original feature vector may be affected by garment deformation; correction improves the accuracy of subsequent comparisons.

[0072] Before making corrections, it is necessary to construct sample feature vectors to obtain the original fabric information and original tensile strength information of the target garment. Based on the original image information, using a method similar to step S202, construct original visual feature vectors representing the stitching, color, and texture of the target garment; based on the original fabric information and original tensile strength information, construct original physical feature vectors representing fiber density and fabric elasticity.

[0073] Based on the weighting factors determined in step S203, the original visual feature matrix and the original physical feature matrix are fused to obtain the sample feature vector. Combining the aligned feature point coordinates, the sample feature vector is corrected, for example, by adjusting certain position-related feature values, ultimately yielding the original feature vector corresponding to the target garment.

[0074] In this embodiment, by analyzing and correcting the deformation of the target garment, the accuracy of the original feature vector is improved, thereby making the subsequent comparison results of the joint feature vector and the original feature vector more reliable, and providing a more solid foundation for accurately assessing the degree of damage to the garment.

[0075] See Figure 4 This is a flowchart illustrating a vision-based clothing insurance loss assessment method provided in Embodiment 4 of this application. Figure 4As shown, after the image acquisition device acquires the current image information of the target garment, the vision-based garment insurance loss assessment method may further include the following steps: Step S401: Based on the semantic segmentation model, perform key region identification on the current image information to determine at least one key region and non-key regions other than the key region.

[0076] Step S402: Differentiate and label the key regions and the non-key regions to obtain the labeling results.

[0077] Step S403: Based on the annotation results, determine the joint feature sub-vector corresponding to the key region from the joint feature vector, and determine the original feature sub-vector corresponding to the key region from the original feature vector corresponding to the target clothing.

[0078] Step S404: Compare the joint feature vector with the original feature vector to determine the comparison result.

[0079] Different areas of clothing have varying degrees of importance in the overall damage assessment. For example, for a shirt, areas like the collar and cuffs are more prone to wear and tear and stains, significantly impacting the garment's value and overall condition, and are therefore considered critical areas. Areas less susceptible to damage, such as the back of the garment, are considered non-critical. Identifying these critical areas allows for more targeted feature analysis and damage assessment.

[0080] Semantic segmentation models can classify and segment different objects or regions in an image. When a current image of a target garment is input into a trained semantic segmentation model, the model will identify at least one key region and non-key regions based on the image's features and pre-learned knowledge. For example, the model might identify the collar, cuffs, and pockets as key regions, while identifying most of the main body of the garment as non-key regions.

[0081] Different annotations are applied to identified key and non-key regions. Annotation methods can be varied, such as using different colors, symbols, or labels. Feature information corresponding to the key regions is extracted from the joint feature vector and the original feature vector to form the joint feature sub-vector and the original feature sub-vector, respectively. This narrows the comparison scope to the key regions. To avoid interference from non-critical areas and improve comparison accuracy, feature elements related to critical areas are selected from the joint feature vector based on the annotation results to form a joint feature sub-vector. Similarly, feature elements corresponding to critical areas are selected from the original feature vector corresponding to the target garment to form the original feature sub-vector. For example, if the joint feature vector contains visual and physical feature information of various areas of the garment, then based on the annotation results, only the feature values ​​corresponding to the critical areas are selected and combined to form the joint feature sub-vector.

[0082] The joint feature vector of the key region is compared with the original feature vector to determine the comparison result of the key region. This comparison result can more accurately reflect the loss of the key region and provide an important basis for the final loss assessment.

[0083] Using a similar comparison method to step S203, such as calculating similarity (cosine similarity, etc.) or distance metrics (Euclidean distance, etc.), the joint feature vector is compared with the original feature vector. A numerical value is obtained from the comparison, representing the degree of similarity or difference between the current state and the original state of the key region; this is the comparison result. For example, the closer the calculated cosine similarity value is to 1, the more similar the current state and the original state of the key region are, and the lower the degree of loss; the closer the value is to 0 or a negative number, the greater the difference and the higher the degree of loss.

[0084] In this embodiment, by identifying key areas, performing differential annotation, extracting feature sub-vectors and comparing them, the focus is on the feature changes of key areas of clothing, thereby improving the accuracy and targeting of clothing insurance loss assessment.

[0085] See Figure 5 This is a flowchart illustrating a vision-based clothing insurance loss assessment method provided in Embodiment 5 of this application. Figure 5 As shown, determining the loss assessment result for the target garment based on the predicted loss level and the corresponding policy information may include the following steps: Step S501: Obtain the insurance policy information of the target garment, extract the insurance information, and determine the policy amount, deductible, deductible level, and depreciation factor.

[0086] Step S502: Determine the loss assessment result for the target garment based on the predicted loss level, the policy amount, the deductible, the deductible level, and the depreciation factor.

[0087] Optionally, step S502 may include the following steps: Detect whether the predicted loss level exceeds the deductible level; If the predicted loss level is detected to exceed the deductible level, the product of the policy amount and the depreciation factor is subtracted from the deductible to obtain the target loss assessment amount, which is the loss assessment result corresponding to the target garment. If the predicted loss level is found to be within the deductible level, the preset loss amount will be used as the loss assessment result.

[0088] The policy information includes various terms and parameters agreed upon between the insurance company and the policyholder. Different policies differ in terms of compensation rules and amounts. By extracting key information from the policy, such as the policy amount, deductible, deductible level, and depreciation factor, basic data is provided for accurate subsequent loss assessment.

[0089] Obtaining the insurance policy information corresponding to the target garment can involve acquiring it from the insurance company's database, electronic policy documents, or other relevant channels. Then, this insurance information is filtered and extracted to identify the policy amount (i.e., the maximum compensation limit stipulated in the insurance contract), deductible (the amount the policyholder bears within a certain range of losses), deductible level (specifying the level of wear and tear under which the insurance company begins to assume liability), and depreciation factor (the percentage reduction in value considering factors such as the garment's usage time and degree of wear).

[0090] First, the predicted damage level is compared with the deductible level stipulated in the policy to determine if it meets the standard for the insurance company to begin compensation. If the predicted damage level exceeds the deductible level, it means that the degree of damage to the clothing has reached the scope of compensation that the insurance company must bear. At this point, the product of the policy amount and the depreciation factor is calculated. This step takes into account the depreciation of the clothing and obtains the actual value of the clothing within the insurance coverage. Then, the deductible is subtracted from this product result, and the difference is the target loss assessment amount, which is the loss assessment result for the target clothing. For example, if the policy amount is 1000 yuan, the depreciation factor is 0.8, and the deductible is 100 yuan, when the damage level exceeds the deductible level, the loss assessment result is 1000 × 0.8 - 100 = 700 yuan.

[0091] If the predicted damage level is within the deductible level, it indicates that the damage to the clothing is minor and does not meet the standard for the insurance company to assume liability. In this case, the pre-set damage amount will be used as the assessment result. The pre-set damage amount is usually a small, predetermined amount or zero, indicating that the insurance company will not provide compensation or will only provide a small amount of compensation in this situation.

[0092] In this embodiment of the application, key policy information is extracted in step S501, and calculations are performed in step S502 based on the relationship between the loss level and the deductible level, as well as other policy parameters, to finally determine the loss assessment result of the target garment, providing a clear basis for the garment insurance claim.

[0093] Corresponding to the vision-based clothing insurance loss assessment method in the above embodiments, Figure 6 This diagram illustrates a structural block diagram of a vision-based clothing insurance loss assessment device provided in Embodiment Six of this application. The aforementioned vision-based clothing insurance loss assessment device can be applied to… Figure 1 The server-side component is shown. For ease of explanation, only the parts relevant to the embodiments of this application are shown.

[0094] See Figure 6 The vision-based clothing insurance claims assessment device includes: The clothing information acquisition module 61 is used to acquire image information of the target clothing based on an image acquisition device, acquire fabric information of the target clothing based on a spectrometer device, acquire tensile strength information of the target clothing based on a tensile sensor, and determine the clothing type corresponding to the target clothing based on the image information. The feature vector analysis module 62 is used to construct visual feature vectors representing the stitches, colors, and textures of the target garment based on the image information, and to construct physical feature vectors representing fiber density and fabric elasticity based on the fabric information and the tensile strength information. The feature comparison module 63 is used to determine a weight factor based on the clothing type, fuse the visual feature matrix and the physical feature matrix based on the weight factor to obtain a joint feature vector, and compare the joint feature vector with the original feature vector corresponding to the target clothing to determine the comparison result. The insurance loss assessment module 64 is used to quantify the comparison results and map them to the loss level to obtain the predicted loss level. Based on the predicted loss level and the corresponding policy information of the target garment, the loss assessment result for the target garment is determined.

[0095] Optionally, the vision-based clothing insurance loss assessment device further includes: The target contour feature determination module is used to determine the target contour features corresponding to at least one target feature region in the target clothing based on the current image information. The mapping matrix determination module is used to acquire the original image information of the target garment, determine the original contour features of the corresponding target feature region from the original image information, and determine the mapping matrix representing the deformation based on the target contour features and the original contour features. The feature point coordinate determination module is used to align the coordinates of each feature point in the original image information of the target garment using the mapping matrix, so as to obtain the aligned feature point coordinates. The original feature vector determination module is used to correct the sample feature vector of the target garment based on the aligned feature point coordinates to obtain the original feature vector corresponding to the target garment.

[0096] Optionally, the vision-based clothing insurance loss assessment device further includes: The clothing information acquisition module is used to acquire the original fabric information and original tensile strength information of the target clothing before correcting the sample feature vector of the target clothing according to the aligned feature point coordinates to obtain the original feature vector corresponding to the target clothing. The construction module is used to construct an original visual feature vector representing the stitches, colors, and textures of the target garment based on the original image information, and to construct an original physical feature vector representing the fiber density and fabric elasticity based on the original fabric information and the original tensile strength information. The matrix fusion module is used to fuse the original visual feature matrix and the original physical feature matrix according to the weighting factor to obtain the sample feature vector.

[0097] Optionally, the feature point coordinate determination module includes: The coordinate mapping unit is used to map and align the coordinates of each feature point in the original image information of the target garment using the moving least squares method based on the mapping matrix, so as to obtain the aligned feature point coordinates.

[0098] Optionally, the vision-based clothing insurance loss assessment device further includes: The non-critical region determination module is used to identify critical regions of the current image information of the target clothing based on the image acquisition device after the current image information is acquired based on the semantic segmentation model, and determine at least one critical region and non-critical regions other than the critical regions. The annotation result acquisition module is used to distinguish and annotate the key regions and the non-key regions to obtain annotation results; The original feature vector determination module is used to determine the joint feature vector corresponding to the key region from the joint feature vector based on the annotation result, and to determine the original feature vector corresponding to the key region from the original feature vector corresponding to the target clothing. The sub-vector comparison module compares the joint feature sub-vector with the original feature sub-vector to determine the comparison result.

[0099] Optionally, the insurance loss assessment module 64 includes: The policy information determination unit is used to obtain the policy information of the target garment, extract the insurance information, and determine the policy amount, deductible, deductible level, and depreciation factor. The damage assessment unit is used to determine the damage assessment result corresponding to the target garment based on the predicted loss level, the policy amount, the deductible, the deductible level, and the depreciation factor.

[0100] Optionally, the damage assessment unit further includes: A level detection subunit is used to detect whether the predicted loss level exceeds the deductible level; The loss assessment result calculation subunit is used to calculate the difference between the product of the policy amount and the depreciation factor and the deductible amount if the predicted loss level is detected to exceed the deductible level, so as to obtain the target loss assessment amount as the loss assessment result corresponding to the target garment. The loss assessment amount determination subunit is used to set a preset loss amount as the loss assessment result if the predicted loss level is detected to be within the deductible level.

[0101] It should be noted that the information interaction and execution process between the above modules, units, and sub-units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0102] Figure 7 This is a schematic diagram of the structure of a computer device provided in Embodiment Seven of this application. Figure 7 As shown, the computer device of this embodiment includes: at least one processor ( Figure 7 The diagram shows only one of the following: a memory and a computer program stored in the memory and executable on at least one processor. When the processor executes the computer program, it implements the steps of any of the above-described vision-based clothing insurance loss assessment methods or embodiments of the vision-based clothing insurance loss assessment method.

[0103] This computer device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that... Figure 7 The examples of computer devices are merely examples and do not constitute a limitation on computer devices. Computer devices may include more or fewer components than shown in the illustration, or combinations of certain components, or different components, such as network interfaces, displays, and input devices.

[0104] The processor referred to can be a CPU, but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0105] Memory includes readable storage media, internal memory, etc., wherein internal memory can be the RAM of a computer device, providing an environment for the operation of the operating system and computer-readable instructions stored in the readable storage media. The readable storage media can be the hard drive of a computer device, or in other embodiments, it can be an external storage device of the computer device, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, memory can include both internal storage units and external storage devices of the computer device. Memory is used to store the operating system, applications, bootloader, data, and other programs, such as program code for computer programs. Memory can also be used to temporarily store data that has been output or will be output.

[0106] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above device can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here. If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the above method embodiments. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. A computer-readable medium can include at least: any entity or device capable of carrying computer program code, a recording medium, a computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0107] The implementation of all or part of the processes in the methods of the above embodiments can also be accomplished by a computer program product. When the computer program product is run on a computer device, it enables the computer device to execute the steps in the above method embodiments.

[0108] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0109] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0110] In the embodiments provided in this application, it should be understood that the disclosed apparatus / computer devices and methods can be implemented in other ways. For example, the apparatus / computer device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0111] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0112] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A vision-based method for assessing losses in clothing insurance, characterized in that, include: The current image information of the target garment is acquired using an image acquisition device, the current fabric information of the target garment is acquired using a spectrometer, and the current tensile strength information of the target garment is acquired using a tensile sensor. Based on the current image information, the garment type corresponding to the target garment is determined. Based on the current image information, a current visual feature vector representing the stitches, color, and texture of the target garment is constructed; based on the current fabric information and the current tensile strength information, a current physical feature vector representing the fiber density and fabric elasticity is constructed. Based on the clothing type, a weighting factor is determined. Based on the weighting factor, the current visual feature matrix and the current physical feature matrix are fused to obtain a joint feature vector. The joint feature vector is then compared with the original feature vector corresponding to the target clothing to determine the comparison result. The comparison results are quantified and mapped to the loss level to obtain the predicted loss level. Based on the predicted loss level and the corresponding policy information of the target garment, the loss assessment result for the target garment is determined.

2. The vision-based clothing insurance loss assessment method according to claim 1, characterized in that, Before comparing the joint feature vector with the original feature vector corresponding to the target garment to determine the comparison result, the method further includes: Based on the current image information, determine the target contour features corresponding to at least one target feature region in the target clothing; Obtain the original image information of the target garment, determine the original contour features of the corresponding target feature region from the original image information, and determine the mapping matrix representing the deformation based on the target contour features and the original contour features; Using the mapping matrix, the coordinates of each feature point in the original image information of the target garment are aligned to obtain the aligned feature point coordinates. Based on the aligned feature point coordinates, the sample feature vector of the target garment is corrected to obtain the original feature vector corresponding to the target garment.

3. The vision-based clothing insurance loss assessment method according to claim 2, characterized in that, Before correcting the sample feature vector of the target garment based on the aligned feature point coordinates to obtain the original feature vector corresponding to the target garment, the method further includes: Obtain the original fabric information and original tensile strength information of the target garment; Based on the original image information, an original visual feature vector representing the stitches, color, and texture of the target garment is constructed; based on the original fabric information and the original tensile strength information, an original physical feature vector representing the fiber density and fabric elasticity is constructed. Based on the weighting factor, the original visual feature matrix and the original physical feature matrix are fused to obtain the sample feature vector.

4. The vision-based clothing insurance loss assessment method according to claim 2, characterized in that, The step of using the mapping matrix to align the coordinates of each feature point in the original image information of the target garment to obtain the aligned feature point coordinates includes: Using the moving least squares method, based on the mapping matrix, the coordinates of each feature point in the original image information of the target garment are mapped and aligned to obtain the aligned feature point coordinates.

5. The vision-based clothing insurance loss assessment method according to claim 1, characterized in that, After acquiring the current image information of the target garment using the image acquisition device, the method further includes: Based on a semantic segmentation model, key regions are identified in the current image information to determine at least one key region and non-key regions other than the key region. The key regions and non-key regions are distinguished and labeled to obtain the labeling results; The step of comparing the joint feature vector with the original feature vector corresponding to the target garment and determining the comparison result includes: Based on the annotation results, a joint feature sub-vector corresponding to the key region is determined from the joint feature vector, and an original feature sub-vector corresponding to the key region is determined from the original feature vector corresponding to the target garment. The joint feature vector is compared with the original feature vector to determine the comparison result.

6. The vision-based clothing insurance loss assessment method according to any one of claims 1 to 5, characterized in that, The step of determining the loss assessment result for the target garment based on the predicted loss level and the corresponding policy information includes: Obtain the insurance policy information of the target garment, extract the insurance information, and determine the policy amount, deductible, deductible level, and depreciation factor; Based on the predicted loss level, the policy amount, the deductible, the deductible level, and the depreciation factor, the loss assessment result corresponding to the target garment is determined.

7. The vision-based clothing insurance loss assessment method according to claim 6, characterized in that, The step of determining the loss assessment result for the target garment based on the predicted loss level, the policy amount, the deductible, the deductible level, and the depreciation factor includes: Detect whether the predicted loss level exceeds the deductible level; If the predicted loss level is detected to exceed the deductible level, the product of the policy amount and the depreciation factor is subtracted from the deductible to obtain the target loss assessment amount, which is the loss assessment result corresponding to the target garment. If the predicted loss level is found to be within the deductible level, the preset loss amount will be used as the loss assessment result.

8. A vision-based clothing insurance loss assessment device, characterized in that, include: The clothing information acquisition module is used to acquire image information of the target clothing based on an image acquisition device, acquire fabric information of the target clothing based on a spectrometer, acquire tensile strength information of the target clothing based on a tensile sensor, and determine the clothing type corresponding to the target clothing based on the image information. The feature vector analysis module is used to construct visual feature vectors representing the stitches, colors, and textures of the target garment based on the image information, and to construct physical feature vectors representing fiber density and fabric elasticity based on the fabric information and tensile strength information. The feature comparison module is used to determine a weighting factor based on the clothing type, fuse the visual feature matrix and the physical feature matrix based on the weighting factor to obtain a joint feature vector, and compare the joint feature vector with the original feature vector corresponding to the target clothing to determine the comparison result. The insurance loss assessment module is used to quantify the comparison results and map them to the loss level to obtain the predicted loss level. Based on the predicted loss level and the corresponding policy information of the target garment, the loss assessment result for the target garment is determined.

9. A computer device, characterized in that, The computer device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the vision-based clothing insurance loss assessment method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the vision-based clothing insurance loss assessment method as described in any one of claims 1 to 7.

Images