Search papers, labs, and topics across Lattice.
This paper introduces BinRVR, a novel binarized framework for RAW video restoration that significantly reduces computation and parameters by approximately 96% while maintaining competitive performance with only a 4% degradation. The framework incorporates a Binarized Information Interaction Module (BIIM) to effectively model both spatial and temporal information, and a Distribution-Aware Binarized Convolution (DAB-Conv) to address quantization errors by leveraging full-precision activation statistics. Extensive experiments validate BinRVR's efficacy in various RAW video restoration tasks and its applicability to downstream video applications like object detection and monocular depth estimation.
Achieving a 96% reduction in computation for RAW video restoration with minimal performance loss opens new avenues for efficient video processing in real-time applications.
RAW video restoration is fundamental to high-quality low-level perception and serves as the basis for a wide range of downstream vision applications. While binary neural networks (BNNs) enable efficient lightweight deployment for image enhancement, their deficiencies in modeling temporal coherence and activation value distributions hinder their effectiveness when applied to video scenarios. In this paper, we propose BinRVR, a binarized RAW video restoration framework that reduces computation and parameters by approximately 96% while incurring only about 4% performance degradation. Specifically, we present a Binarized Information Interaction Module (BIIM) to jointly model spatial and temporal information in an efficient and unified manner. Moreover, we develop a Distribution-Aware Binarized Convolution (DAB-Conv) that leverages the statistics of full-precision activations to mitigate quantization errors. The proposed framework further supports multi-bit quantization, enabling flexible accuracy-efficiency trade-offs across different hardware constraints. Extensive experiments demonstrate that our BinRVR achieves competitive performance compared with state-of-the-art binarized methods on RAW video restoration tasks, including low-light enhancement, denoising, deblurring, and super-resolution. We further explore the potential of our method on downstream video applications, including object detection and monocular depth estimation.
