A target detection model training method, device, electronic equipment and storage medium

Abstract

The present application provides a target detection model training method, device, electronic equipment and storage medium. The method includes: obtaining multiple sample images and multiple categories of thermal images; using multiple sample images as training data, and using multiple categories of thermal images As the data label, train the feature extraction network and classification attention network in the target detection model, and use the labeled image as the data label to train the feature extraction network and detection network in the target detection model, and obtain The trained object detection model. By adding a classification attention network that can extract and effectively utilize the differentiated features between different categories in the target detection model, it guides the model to locate key areas that have important influences in the image to be detected, allowing the model to better focus on key areas. In this way, the difference between confusing categories can be learned, and the correct rate of target detection for images can be improved.

Description

technical field [0001] The present application relates to the technical field of target detection, in particular, to a target detection model training method, device, electronic equipment and storage medium. Background technique [0002] Target detection, also known as target extraction, is an image understanding algorithm based on the geometric and statistical characteristics of the target. Target detection combines target positioning and recognition. For example, based on computer vision algorithms, it detects The target of interest, that is, the position of the target is marked with a rectangular frame, and the category of the target is identified. [0003] In the current target detection algorithm, a difficult problem recognized in the industry is the contradiction between target positioning and target classification. Target positioning emphasizes the spatial relationship between features, while target classification emphasizes the spatial invariance of features. Therefo...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the embodiments of the present application is to provide a target detection model training method, device, electronic equipment, and storage medium, which are used to improve the problem of low accuracy of target detection in images under certain circumstances

Images