English

Compressed Sensing of Generative Sparse-latent (GSL) Signals

Machine Learning 2023-10-24 v1 Artificial Intelligence

Abstract

We consider reconstruction of an ambient signal in a compressed sensing (CS) setup where the ambient signal has a neural network based generative model. The generative model has a sparse-latent input and we refer to the generated ambient signal as generative sparse-latent signal (GSL). The proposed sparsity inducing reconstruction algorithm is inherently non-convex, and we show that a gradient based search provides a good reconstruction performance. We evaluate our proposed algorithm using simulated data.

Keywords

Cite

@article{arxiv.2310.15119,
  title  = {Compressed Sensing of Generative Sparse-latent (GSL) Signals},
  author = {Antoine Honoré and Anubhab Ghosh and Saikat Chatterjee},
  journal= {arXiv preprint arXiv:2310.15119},
  year   = {2023}
}

Comments

Accepted at 31st European Signal Processing Conference, EUSIPCO 2023