Image annotation and element extraction method and system for anti-fraud vehicle insurance

Abstract

The invention discloses an image annotation and element extraction method and system for vehicle insurance anti-fraud. According to the method, anti-fraud element extraction is carried out on images such as vehicle insurance site collection and post supplementary pictures. The system comprises a vehicle insurance element table construction module, an image acquisition module, an annotation function module and an element extraction module, wherein the annotation function module comprises a multi-label category annotation module, a vehicle damage part annotation module and a face annotation module; and the element extraction module is used for performing element extraction on each annotation data set. The method mainly focuses on establishment of image element annotation and extraction facing vehicle insurance fraud prevention, so that the extracted image elements are more objective, vehicle insurance structured data which can be used for cross validation are generated, and the data quality is improved.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to an image labeling and element extraction method and system for auto insurance anti-fraud. Background technique [0002] At the same time, with the development of informatization in the financial and insurance industry, the relevant business data is growing rapidly. How to use the rapidly growing data, especially the more objective image data, to detect insurance fraud and effectively combat and deter anti-fraud, has a great impact on the auto insurance industry. have important meaning. The application of the existing intelligent recognition technology in the vehicle insurance industry is mostly aimed at the insurance loss determination. For example, Chinese patent CN113344712A discloses an intelligent sorting and insurance compensation system based on image recognition, and Chinese patent CN113706513A discloses a car damage image based on image detection. an...

AI Technical Summary

Problems solved by technology

Secondly, the existing image models mostly use public databases, and there are few types of elements that can be extracted, so they have little application value in anti-fraud

Thirdly, the image data labeling of car insurance is not targeted. Usually, only a small amount of car damage features are added when the pre-training model is fine-tuned, so that the extraction results contain a lot of noise features, which affects the judgment of the anti-fraud model

Images