Basis Prediction Networks for Effective Burst Denoising with Large Kernels
This addresses the problem of efficient and high-quality burst denoising for image processing applications, representing an incremental advancement by optimizing kernel representation.
The paper tackles burst image denoising by introducing a basis prediction network that reduces output dimensionality, enabling the use of larger kernels and achieving over 1dB PSNR improvement with faster run-times compared to state-of-the-art methods.
Bursts of images exhibit significant self-similarity across both time and space. This motivates a representation of the kernels as linear combinations of a small set of basis elements. To this end, we introduce a novel basis prediction network that, given an input burst, predicts a set of global basis kernels -- shared within the image -- and the corresponding mixing coefficients -- which are specific to individual pixels. Compared to state-of-the-art techniques that output a large tensor of per-pixel spatiotemporal kernels, our formulation substantially reduces the dimensionality of the network output. This allows us to effectively exploit comparatively larger denoising kernels, achieving both significant quality improvements (over 1dB PSNR) and faster run-times over state-of-the-art methods.