Efficient Blind-Spot Neural Network Architecture for Image Denoising
This addresses the challenge of training denoising models when clean images are unavailable, which is a common issue in computational photography, though it appears incremental as it builds on prior blind-spot methods.
The paper tackles the problem of image denoising without clean training data by proposing a novel blind-spot neural network architecture using dilations, achieving state-of-the-art results on established datasets.
Image denoising is an essential tool in computational photography. Standard denoising techniques, which use deep neural networks at their core, require pairs of clean and noisy images for its training. If we do not possess the clean samples, we can use blind-spot neural network architectures, which estimate the pixel value based on the neighbouring pixels only. These networks thus allow training on noisy images directly, as they by-design avoid trivial solutions. Nowadays, the blind-spot is mostly achieved using shifted convolutions or serialization. We propose a novel fully convolutional network architecture that uses dilations to achieve the blind-spot property. Our network improves the performance over the prior work and achieves state-of-the-art results on established datasets.