当前位置：网站首页>CVPR 2022 oral | Dalian University of technology proposes to identify deepbdc with small samples, and the performance of six benchmarks is the best

CVPR 2022 oral | Dalian University of technology proposes to identify deepbdc with small samples, and the performance of six benchmarks is the best

2022-04-22 13:54:00 【Zhiyuan community】

In this paper , Researchers from Dalian University of technology and other institutions put forward DeepBDC For small sample classification tasks ,DeepBDC By measuring the joint distribution between sample pairs, we can obtain more accurate similarity , It greatly improves the performance of small sample classification . The paper has been CVPR 2022 Received as Oral Presentation.

In this paper, Brownian distance covariance, a similarity measure based on probability and statistics, is introduced into deep learning for the first time , An end-to-end small sample recognition method is proposed DeepBDC. The proposed method includes general target recognition 、 Fine-grained classification and cross domain classification 6 The best performance has been achieved on all standard databases . The paper has been CVPR 2022 Received as Oral Presentation.

版权声明
本文为[Zhiyuan community]所创，转载请带上原文链接，感谢
https://yzsam.com/2022/04/202204221349022504.html

边栏推荐

猜你喜欢

随机推荐