Model training method, model training device, storage medium and electronic equipment

Abstract

The embodiment of the invention discloses a model training method, a model training device, a storage medium and electronic equipment. According to the embodiment of the invention, the method comprises steps of after a to-be-trained model of a source domain, a first image set and a second image set of a target domain are obtained, training the to-be-trained model by taking each first image as an input and taking an object recognition frame corresponding to each first image and the category of the object recognition frame as training targets, meanwhile, taking each second image as an input, obtaining outputs of the second images in different layers of the to-be-trained model so as to train the attention discrimination model and the feature discrimination model, thereby responding to convergence of a detection loss function of the to-be-trained model, an attention loss function of the attention discrimination model and a feature loss function of the feature discrimination model, and obtaining a training result;determining the trained to-be-trained model as a target model. According to the embodiment of the invention, the to-be-trained model can be trained without marking the target domain image, the object detection capability of the model for the target domain is improved, and the marking cost is reduced.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a model training method, a model training device, a storage medium and electronic equipment. Background technique [0002] In the field of traditional vision, object detection is a very popular research direction. With the continuous development of science and technology, people can perform more accurate object detection on the target object in the RGB pixel matrix (that is, the color image collected by the image acquisition device) by means of machine learning. However, the RGB pixel matrix is ​​easily affected by the acquisition environment, such as the light intensity and visibility of the external environment. Therefore, when performing object detection on the RGB pixel matrix obtained in different environments, there will be certain limitations in the way of machine learning. For The detection accuracy of the RGB pixel matrix obtained in some environments is high, w...

AI Technical Summary

Problems solved by technology

However, the RGB pixel matrix is ​​easily affected by the acquisition environment, such as the light intensity and visibility of the external environment. Therefore, when performing object detection on the RGB pixel matrix obtained in different environments, there will be certain limitations in the way of machine learning. For The detection accuracy of the RGB pixel matrix obtained in some environments is high, while the detection accuracy of the RGB pixel matrix obtained in other environments is not high

Images