Dual temporal memory network with high-order spatio-temporal graph learning for video object segmentation

Jiaqing Fan, Shenglong Hu, Long Wang, Kaihua Zhang, Bo Liu

Abstract: Highlights•We develop DTM-VOS model to keep both long- and short-term consistency.•We propose high-order graph model to simulate near-instantaneous interactions.•We design a S-GRU module to represent object-oriented features.•We present a direction-aware attention mechanism.