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: Due to wide field of view and background confusion, remote sensing objects are small and densely packed, hence commonly used detection methods detecting small objects are not satisfactory.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results