Abstract: How to identify and segment camouflaged objects from the background is challenging. Inspired by the multi-head self-attention in Transformers, we present a simple masked separable attention ...
Abstract: Deep learning techniques continue to be used in various applications in recent years. However, when it is difficult to obtain adequate training samples, the performance of the depth model ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results