Hybrid Training of Denoising Networks to Improve the Texture Acutance of Digital Cameras
This work addresses texture rendering issues in digital cameras, offering a systematic improvement for real imaging devices, though it appears incremental in its approach.
The paper tackled the problem of improving texture acutance in digital cameras by proposing a hybrid training method for denoising networks, resulting in strong enhancement of the acutance metric without compromising fidelity, as demonstrated on RGB and RAW image processing.
In order to evaluate the capacity of a camera to render textures properly, the standard practice, used by classical scoring protocols, is to compute the frequential response to a dead leaves image target, from which is built a texture acutance metric. In this work, we propose a mixed training procedure for image restoration neural networks, relying on both natural and synthetic images, that yields a strong improvement of this acutance metric without impairing fidelity terms. The feasibility of the approach is demonstrated both on the denoising of RGB images and the full development of RAW images, opening the path to a systematic improvement of the texture acutance of real imaging devices.