Multi-scale Convolutional Neural Networks for Inverse Problems

Abstract

Inverse problems exist in many domains such as phase imaging, image processing, and computer vision. These problems are often solved with application-specific algorithms, even though their nature remains the same: mapping input image(s) to output image(s). Deep convolutional neural networks have shown great potential for highly variable tasks across many image-based domains, but are usually difficult to train due to their inner high non-linearities. We propose a novel neural network architecture highlighting fast convergence as a generic solution addressing image(s)-to-image(s) inverse problems of different domains. Here we show that this approach is effective at predicting phases from direct intensity measurements, imaging objects from diffused reflections and denoising scanning transmission electron microscopy images, with just different training datasets. This opens a way to solve problems statistically through big data, in contrast to implementing explicit inversion algorithms from their mathematical formulas. Previous works have targeted much more on \textit{how} can we reconstruct rather than \textit{what} can be reconstructed. Our strategy offers a paradigm shift.

13 pages, 7 figures

References in corpus (18)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising
• Large Scale GAN Training for High Fidelity Natural Image Synthesis
• Deep Convolutional Neural Network for Inverse Problems in Imaging
• XLNet: Generalized Autoregressive Pretraining for Language Understanding
• Progressive Growing of GANs for Improved Quality, Stability, and Variation
• Phase recovery and holographic image reconstruction using deep learning in neural networks
• A disciplined approach to neural network hyper-parameters: Part 1 -- learning rate, batch size, momentum, and weight decay
• Neural Architecture Optimization
• SAR Image Despeckling Using a Convolutional Neural Network
• Wide Neural Networks of Any Depth Evolve as Linear Models Under Gradient Descent
• Fast Automated Analysis of Strong Gravitational Lenses with Convolutional Neural Networks
• Low Photon Count Phase Retrieval Using Deep Learning
• Optically lightweight tracking of objects around a corner
• The Expressive Power of Neural Networks: A View from the Width
• Real-time coherent diffraction inversion using deep generative networks
• Auto-Keras: An Efficient Neural Architecture Search System
• Transport of Intensity Phase Retrieval of Arbitrary Wave Fields Including Vortices

Cited by in corpus (14)

• Probabilistic Deep Learning for Real-Time Large Deformation Simulations
• Theory of the Frequency Principle for General Deep Neural Networks
• MAgNET: A Graph U-Net Architecture for Mesh-Based Simulations
• Exploring particle dynamics during self-organization processes via rotationally invariant latent representations
• Inverting the Kohn-Sham equations with physics-informed machine learning
• Deep frequency principle towards understanding why deeper learning is faster
• Learning Self-Prior for Mesh Inpainting Using Self-Supervised Graph Convolutional Networks
• Self-supervised physics-informed generative networks for phase retrieval from a single X-ray hologram
• Scaling Up Bayesian Uncertainty Quantification for Inverse Problems using Deep Neural Networks
• Edge Sparse Basis Network: A Deep Learning Framework for EEG Source Localization
• Multi-resolution Physics-Aware Recurrent Convolutional Neural Network for Complex Flows
• Self-supervised denoising for massive noisy images
• Fourier-domain Variational Formulation and Its Well-posedness for Supervised Learning
• PlumeCityNet: Multi-Resolution Air Quality Forecasting