A Decoding Scheme with Successive Aggregation of Multi-Level Features for Light-Weight Semantic Segmentation

arXiv:2402.11201v1 Announce Type: new
Abstract: Multi-scale architecture, including hierarchical vision transformer, has been commonly applied to high-resolution semantic segmentation to deal with computational complexity with minimum performance loss. In this paper, we propose a novel decoding scheme for semantic segmentation in this regard, which takes multi-level features from the encoder with multi-scale architecture. The decoding scheme based on a multi-level vision transformer aims to achieve not only reduced computational expense but also higher segmentation accuracy, by introducing successive cross-attention in …

abstract aggregation architecture arxiv complexity computational cs.cv deal decoding features hierarchical light loss novel paper performance regard scale segmentation semantic transformer type vision