Model acquisition method, model acquisition device and intelligent equipment

Abstract

The invention discloses a model acquisition method, a model acquisition device, intelligent equipment and a computer readable storage medium. The method comprises the following steps: encoding a sample image through a to-be-trained feature extraction encoder; decoding the encoding result through a to-be-trained edge detection decoder and a to-be-trained portrait segmentation decoder to obtain an edge result of the sample image outputted by the edge detection decoder and a segmentation result of the sample image outputted by the portrait segmentation decoder; calculating total loss based on theedge result, the segmentation result and the sample image; and optimizing the edge detection decoder, the portrait segmentation decoder and the feature extraction encoder according to the total loss,and obtaining a trained portrait segmentation model formed by the trained feature extraction encoder and the trained portrait segmentation decoder after the total loss reaches convergence. Through the scheme of the invention, the portrait segmentation result of the portrait segmentation model can be more accurate.

Description

technical field [0001] The present application belongs to the technical field of image processing, and in particular relates to a model acquisition method, a model acquisition device, an intelligent device, and a computer-readable storage medium. Background technique [0002] In recent years, with the rapid development of artificial intelligence (AI), many AI-based applications have been developed. These AI applications can be deployed in robots to interact with users through AI algorithms. In the process of AI applications interacting with users, it is particularly important to perform accurate portrait segmentation on users. [0003] At present, the common portrait segmentation algorithm based on fully connected network (FCN) can usually roughly segment the outline of the portrait. However, when the user interacts with the AI ​​application, the user's environment is ever-changing; when the difference between the user and the environment is not very large, these portrait ...

Images