ColorMNet: A Memory-based Deep Spatial-Temporal Feature Propagation Network for Video Colorization

arXiv:2404.06251v1 Announce Type: new
Abstract: How to effectively explore spatial-temporal features is important for video colorization. Instead of stacking multiple frames along the temporal dimension or recurrently propagating estimated features that will accumulate errors or cannot explore information from far-apart frames, we develop a memory-based feature propagation module that can establish reliable connections with features from far-apart frames and alleviate the influence of inaccurately estimated features. To extract better features from each frame for the above-mentioned feature propagation, we explore …

arxiv colorization cs.cv feature memory network propagation spatial temporal type video