Image Reconstruction with Predictive Filter Flow
This addresses image quality enhancement for computer vision applications, offering an interpretable alternative to black-box models.
The authors tackled image reconstruction problems like denoising and deconvolution by proposing a framework that synthesizes spatially varying linear filters to process corrupted input images, achieving substantial performance improvements over state-of-the-art methods across tasks such as motion blur removal, artifact reduction, and super-resolution while being significantly faster than optimization-based approaches.
We propose a simple, interpretable framework for solving a wide range of image reconstruction problems such as denoising and deconvolution. Given a corrupted input image, the model synthesizes a spatially varying linear filter which, when applied to the input image, reconstructs the desired output. The model parameters are learned using supervised or self-supervised training. We test this model on three tasks: non-uniform motion blur removal, lossy-compression artifact reduction and single image super resolution. We demonstrate that our model substantially outperforms state-of-the-art methods on all these tasks and is significantly faster than optimization-based approaches to deconvolution. Unlike models that directly predict output pixel values, the predicted filter flow is controllable and interpretable, which we demonstrate by visualizing the space of predicted filters for different tasks.