Temporal Memory Attention for Video Semantic Segmentation

Abstract

Video semantic segmentation requires to utilize the complex temporal relations between frames of the video sequence. Previous works usually exploit accurate optical flow to leverage the temporal relations, which suffer much from heavy computational cost. In this paper, we propose a Temporal Memory Attention Network (TMANet) to adaptively integrate the long-range temporal relations over the video sequence based on the self-attention mechanism without exhaustive optical flow prediction. Specially, we construct a memory using several past frames to store the temporal information of the current frame. We then propose a temporal memory attention module to capture the relation between the current frame and the memory to enhance the representation of the current frame. Our method achieves new state-of-the-art performances on two challenging video semantic segmentation datasets, particularly 80.3% mIoU on Cityscapes and 76.5% mIoU on CamVid with ResNet-50.

arXiv commentICIP 2021

References in corpus (2)

• Efficient Semantic Video Segmentation with Per-frame Inference
• Fast Video Object Segmentation using the Global Context Module

Cited by in corpus (6)

• Spatio-Temporal Self-Attention Network for Video Saliency Prediction
• UVid-Net: Enhanced Semantic Segmentation of UAV Aerial Videos by Embedding Temporal Information
• Semantic Segmentation on VSPW Dataset through Aggregation of Transformer Models
• Memory Based Video Scene Parsing
• TBN-ViT: Temporal Bilateral Network with Vision Transformer for Video Scene Parsing
• Scene-aware Learning Network for Radar Object Detection