Model training, target detection and rendering method, storage medium, program product

Abstract

The application provides a model training, target detection and rendering method, a storage medium, a program product, and a training method of an image enhancement model, which comprises the following steps: obtaining a sample image of a target scene under a first light environment and a labeled image under a second light environment, wherein the light intensity of the first light environment is smaller than that of the second light environment; performing light enhancement processing on the sample image by using an image enhancement model to obtain a light enhanced image; calculating a pair loss value according to a first difference value between the light enhanced image and the sample image and a second difference value between the light enhanced image and the labeled image, wherein the pair loss value is inversely proportional to the first difference value and proportional to the second difference value; and training the image enhancement model according to the pair loss value. The scheme can improve the quality of the light enhanced image from the perspective of machine vision, and when the light enhanced image is applied to a downstream task such as target detection, the accuracy of the downstream task can be improved.

Images