Superdense-scale network for semantic segmentation

Zhiqiang Li, Jie Jiang, Xi Chen, Honggang Qi, Qingli Li, Jiapeng Liu, Laiwen Zheng, Min Liu, Yundong Zhang

Abstract: Highlights•PSSAC exploits superdense-scale semantics and extends ASPP at dense scales.•PSSAC increases receptive fields for larger-scale variations at little extra cost.•Extensive experiments proves the superiority of SDSNet.