Algorithmic Guarantees for Inverse Imaging with Untrained Network Priors

Deep neural networks as image priors have been recently introduced for problems such as denoising, super-resolution and inpainting with promising performance gains over hand-crafted image priors such as sparsity and low-rank. Unlike learned generative priors they do not require any training over large datasets. However, few theoretical guarantees exist in the scope of using untrained neural network priors for inverse imaging problems. We explore new applications and theory for untrained neural network priors. Specifically, we consider the problem of solving linear inverse problems, such as compressive sensing, as well as non-linear problems, such as compressive phase retrieval. We model images to lie in the range of an untrained deep generative network with a fixed seed. We further present a projected gradient descent scheme that can be used for both compressive sensing and phase retrieval and provide rigorous theoretical guarantees for its convergence. We also show both theoretically as well as empirically that with deep network priors, one can achieve better compression rates for the same image quality compared to hand crafted priors.

[1]  Pascal Vincent,et al.  Stacked Denoising Autoencoders: Learning Useful Representations in a Deep Network with a Local Denoising Criterion , 2010, J. Mach. Learn. Res..

[2]  Chinmay Hegde,et al.  Sample-Efficient Algorithms for Recovering Structured Signals From Magnitude-Only Measurements , 2017, IEEE Transactions on Information Theory.

[3]  Vladislav Voroninski,et al.  Phase Retrieval Under a Generative Prior , 2018, NeurIPS.

[4]  Tamás Sarlós,et al.  Improved Approximation Algorithms for Large Matrices via Random Projections , 2006, 2006 47th Annual IEEE Symposium on Foundations of Computer Science (FOCS'06).

[5]  Barnabás Póczos,et al.  Gradient Descent Provably Optimizes Over-parameterized Neural Networks , 2018, ICLR.

[6]  A. Bruckstein,et al.  K-SVD : An Algorithm for Designing of Overcomplete Dictionaries for Sparse Representation , 2005 .

[7]  Michael A. Saunders,et al.  Atomic Decomposition by Basis Pursuit , 1998, SIAM J. Sci. Comput..

[8]  Alexandros G. Dimakis,et al.  Compressed Sensing with Deep Image Prior and Learned Regularization , 2018, ArXiv.

[9]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[10]  Yingbin Liang,et al.  Reshaped Wirtinger Flow for Solving Quadratic System of Equations , 2016, NIPS.

[11]  Reinhard Heckel,et al.  Deep Decoder: Concise Image Representations from Untrained Non-convolutional Networks , 2018, ICLR.

[12]  Yuxin Chen,et al.  Solving Random Quadratic Systems of Equations Is Nearly as Easy as Solving Linear Systems , 2015, NIPS.

[13]  S. Mallat.  A wavelet tour of signal processing , 1998 .

[14]  Samet Oymak,et al.  Toward Moderate Overparameterization: Global Convergence Guarantees for Training Shallow Neural Networks , 2019, IEEE Journal on Selected Areas in Information Theory.

[15]  Lior Wolf,et al.  Audio Denoising with Deep Network Priors , 2019, ArXiv.

[16]  Alexandros G. Dimakis,et al.  One-dimensional Deep Image Prior for Time Series Inverse Problems , 2019, 2022 56th Asilomar Conference on Signals, Systems, and Computers.

[17]  Chinmay Hegde,et al.  Alternating Phase Projected Gradient Descent with Generative Priors for Solving Compressive Phase Retrieval , 2019, ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[18]  Richard G. Baraniuk,et al.  prDeep: Robust Phase Retrieval with a Flexible Deep Network , 2018, ICML.

[19]  Volkan Cevher,et al.  Model-Based Compressive Sensing , 2008, IEEE Transactions on Information Theory.

[20]  Chinmay Hegde,et al.  Solving Linear Inverse Problems Using Gan Priors: An Algorithm with Provable Guarantees , 2018, 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[21]  Karen O. Egiazarian,et al.  Image denoising with block-matching and 3D filtering , 2006, Electronic Imaging.

[22]  Gang Wang,et al.  Sparse Phase Retrieval via Truncated Amplitude Flow , 2016, IEEE Transactions on Signal Processing.

[23]  S. Frick,et al.  Compressed Sensing , 2014, Computer Vision, A Reference Guide.

[24]  Michael Elad,et al.  Multilayer Convolutional Sparse Modeling: Pursuit and Dictionary Learning , 2017, IEEE Transactions on Signal Processing.

[25]  Xiaoou Tang,et al.  Image Super-Resolution Using Deep Convolutional Networks , 2014, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[26]  Alexandros G. Dimakis,et al.  Compressed Sensing using Generative Models , 2017, ICML.

[27]  Chinmay Hegde,et al.  Fast, Sample-Efficient Algorithms for Structured Phase Retrieval , 2017, NIPS.

[28]  Michael Elad,et al.  Convolutional Dictionary Learning via Local Processing , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[29]  Reinhard Heckel,et al.  Regularizing linear inverse problems with convolutional neural networks , 2019, ArXiv.

[30]  Soumith Chintala,et al.  Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks , 2015, ICLR.

[31]  Andrea Vedaldi,et al.  Deep Image Prior , 2017, International Journal of Computer Vision.

[32]  Yin Zhang,et al.  User’s Guide for TVAL3: TV Minimization by Augmented Lagrangian and Alternating Direction Algorithms , 2010 .

[33]  M. Elad,et al.  $rm K$-SVD: An Algorithm for Designing Overcomplete Dictionaries for Sparse Representation , 2006, IEEE Transactions on Signal Processing.

[34]  Timothy Lillicrap,et al.  Deep Compressed Sensing , 2019, ICML.

[35]  Deanna Needell,et al.  CoSaMP: Iterative signal recovery from incomplete and inaccurate samples , 2008, ArXiv.

[36]  Chun-Liang Li,et al.  One Network to Solve Them All — Solving Linear Inverse Problems Using Deep Projection Models , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[37]  Ali Ahmed,et al.  Deep Ptych: Subsampled Fourier Ptychography Using Generative Priors , 2018, ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[38]  Xiaodong Li,et al.  Optimal Rates of Convergence for Noisy Sparse Phase Retrieval via Thresholded Wirtinger Flow , 2015, ArXiv.