Spatiotemporal Blind-Spot Network with Calibrated Flow Alignment for Self-Supervised Video Denoising

• DOI: 10.1609/aaai.v39i3.32242
• Publication Date: 2025-04-11T09:47:17Z
• Authors: Zikang Chen; Tao Jiang; Xiaowan Hu; Wang Zhang; Huaqiu Li; Haoqian Wang

Self-supervised video denoising aims to remove noise from videos without relying on ground truth data, leveraging the itself recover clean frames. Existing methods often rely simplistic feature stacking or apply optical flow thorough analysis. This results in suboptimal utilization of both inter-frame and intra-frame information, it also neglects potential alignment under self-supervised conditions, leading biased insufficient outcomes. To this end, we first explore practicality setting introduce a SpatioTemporal Blind-spot Network (STBN) for global frame utilization. In temporal domain, utilize bidirectional blind-spot propagation through proposed block ensure accurate effectively capture long-range dependencies. spatial receptive field expansion module, which enhances improves perception capabilities. Additionally, reduce sensitivity estimation noise, propose an unsupervised distillation mechanism that refines fine-grained interactions during alignment. Our method demonstrates superior performance across synthetic real-world datasets.