Search papers, labs, and topics across Lattice.
This paper introduces EvTexture++, an innovative event-driven framework that enhances texture recovery in video super-resolution (VSR) by leveraging high-frequency spatiotemporal details from event signals. The method incorporates a specialized texture enhancement branch and an iterative module that progressively refines texture regions, addressing both intra-frame recovery and inter-frame temporal consistency to prevent flickering. Experimental results show that EvTexture++ achieves state-of-the-art performance, significantly improving existing VSR models with gains of up to 1.55 dB in PSNR on the Vid4 dataset.
Texture recovery in video super-resolution just got a major upgrade, with EvTexture++ achieving unprecedented detail and consistency through event-driven enhancements.
Event-based vision has drawn increasing attention owing to its distinctive properties, including ultra-high temporal resolution and extreme dynamic range. Recent works have introduced it to video super-resolution (VSR) to enhance flow estimation and temporal alignment. In contrast, this paper shifts the focus of event signals from motion refinement to texture enhancement in VSR. We propose EvTexture++, the first event-driven framework dedicated to texture enhancement in VSR. It leverages high-frequency spatiotemporal details from events to improve texture recovery. EvTexture++ incorporates a customized texture enhancement branch, along with an iterative texture enhancement module that progressively exploits high-temporal-resolution event information for texture restoration. This enables gradual refinement of texture regions across iterations, yielding more accurate and detailed high-resolution outputs. Besides intra-frame texture recovery, large motions could degrade inter-frame temporal consistency, particularly in texture regions, leading to texture flickering. To mitigate this, we further exploit the continuous-time motion cues of events to enhance temporal consistency, introducing a temporal texture alignment module that estimates event-guided texture-aware flow for precise inter-frame texture alignment. Moreover, EvTexture++ is designed as a plug-and-play tool to flexibly boost the performance of existing VSR models. Experiments on five datasets demonstrate that EvTexture++ achieves state-of-the-art performance. When integrated into recent VSR models, it yields significant improvements, with gains of up to 1.55 dB in PSNR on the texture-rich Vid4 dataset.
