Training Implicit Networks for Image Deblurring using Jacobian-Free Backpropagation
This work addresses the computational bottleneck in training implicit networks for inverse imaging problems, offering a more efficient method for researchers and practitioners in computer vision.
The paper tackled the challenge of training implicit networks for image deblurring by applying Jacobian-Free Backpropagation (JFB) to avoid expensive gradient calculations, achieving results comparable to fine-tuned optimization schemes and state-of-the-art networks at reduced computational cost.
Recent efforts in applying implicit networks to solve inverse problems in imaging have achieved competitive or even superior results when compared to feedforward networks. These implicit networks only require constant memory during backpropagation, regardless of the number of layers. However, they are not necessarily easy to train. Gradient calculations are computationally expensive because they require backpropagating through a fixed point. In particular, this process requires solving a large linear system whose size is determined by the number of features in the fixed point iteration. This paper explores a recently proposed method, Jacobian-free Backpropagation (JFB), a backpropagation scheme that circumvents such calculation, in the context of image deblurring problems. Our results show that JFB is comparable against fine-tuned optimization schemes, state-of-the-art (SOTA) feedforward networks, and existing implicit networks at a reduced computational cost.