On the unreasonable effectiveness of CNNs

2020/07/29 by Andreas Hauptmann, Jonas Adler, Hauptmann, Andreas +1 · 2 voices
Computer Science · Engineering · #Advanced X-ray and CT Imaging #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #FOS: Electrical engineering #Image and Video Processing (eess.IV) #Medical Imaging and Analysis #Neural and Evolutionary Computing (cs.NE) #cs.CV #cs.NE #eess.IV #electronic engineering #information engineering

paper · pdf · doi:10.48550/arxiv.2007.14745

openalex publication_date 2020/07/29 · arxiv published 2020/07/29 · arxiv updated 2020/07/29 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Deep learning methods using convolutional neural networks (CNN) have been successfully applied to virtually all imaging problems, and particularly in image reconstruction tasks with ill-posed and complicated imaging models. In an attempt to put upper bounds on the capability of baseline CNNs for solving image-to-image problems we applied a widely used standard off-the-shelf network architecture (U-Net) to the "inverse problem" of XOR decryption from noisy data and show acceptable results.

Discussions

• On the Unreasonable Effectiveness of CNNs [hn, 2 points, 1 comments]
• On the Unreasonable Effectiveness of CNNs [hn, 1 points, 0 comments]

Related

• Deep Convolutional Neural Network for Inverse Problems in Imaging
• Deep Learning Techniques for Inverse Problems in Imaging
• When to Use Convolutional Neural Networks for Inverse Problems
• Improving Robustness of Deep-Learning-Based Image Reconstruction
• Return of the Devil in the Details: Delving Deep into Convolutional Nets