Beyond Deep Residual Learning for Image Restoration: Persistent Homology-Guided Manifold Simplification
This work addresses performance limitations in image restoration for complex images, offering a faster and more effective solution for applications like denoising and super-resolution.
The paper tackles the problem of image restoration when images contain complex patterns and structures, where existing CNNs perform poorly, by proposing a feature space deep residual learning algorithm that outperforms state-of-the-art methods in denoising and super-resolution tasks, achieving third place in the NTIRE competition with 5-10 times faster computational time.
The latest deep learning approaches perform better than the state-of-the-art signal processing approaches in various image restoration tasks. However, if an image contains many patterns and structures, the performance of these CNNs is still inferior. To address this issue, here we propose a novel feature space deep residual learning algorithm that outperforms the existing residual learning. The main idea is originated from the observation that the performance of a learning algorithm can be improved if the input and/or label manifolds can be made topologically simpler by an analytic mapping to a feature space. Our extensive numerical studies using denoising experiments and NTIRE single-image super-resolution (SISR) competition demonstrate that the proposed feature space residual learning outperforms the existing state-of-the-art approaches. Moreover, our algorithm was ranked third in NTIRE competition with 5-10 times faster computational time compared to the top ranked teams. The source code is available on page : https://github.com/iorism/CNN.git