Progressive compressive sensing of large images with multiscale deep learning reconstruction

Abstract

Compressive sensing (CS) is a sub-Nyquist sampling framework that has been employed to improve the performance of numerous imaging applications during the last 15 years. Yet, its application for large and high-resolution imaging remains challenging in terms of the computation and acquisition effort involved. Often, low-resolution imaging is sufficient for most of the considered tasks and only a fraction of cases demand high resolution, but the problem is that the user does not know in advance when high-resolution acquisition is required. To address this, we propose a multiscale progressive CS method for the high-resolution imaging. The progressive sampling refines the resolution of the image, while incorporating the already sampled low-resolution information, making the process highly efficient. Moreover, the multiscale property of the progressively sensed samples is capitalized for a fast, deep learning (DL) reconstruction, otherwise infeasible due to practical limitations of training on high-resolution images. The progressive CS and the multiscale reconstruction method are analyzed numerically and demonstrated experimentally with a single pixel camera imaging system. We demonstrate 4-megapixel size progressive compressive imaging with about half the overall number of samples, more than an order of magnitude faster reconstruction, and improved reconstruction quality compared to alternative conventional CS approaches.

Comparison of deep learning-based compressive imaging from a practitioner's viewpoint

“…3 and 4, the extent of dependency on the Gaussian and quantization noise depends on the CI method and compression ratio. In general, CMS-Net [11] outperforms the other methods provided that the sensing is done with at least 8 bits. In such a case it exhibits the best performance for SNRs higher than 30dB at 10% compression ratio, and at CR=1% with SNRs higher than 15 dB.…”

“…Out of its four versions, we tested AMP_Net_K, over 6 phases and 9 phases. [11]. This model uses the Hadamard transform by taking the smallest block transformed, from the transformed image, reconstructing it, and adding the reconstruction to the bigger block in size until we receive the reconstructed image.…”

Reducing ringing artefact in fresnel digital holography using compressed sensing

“…24,25 Recently, machine learning methods have also been proposed for finding the CS reconstruction. 26,27 Next, we outline the OMP algorithm. In a minimum L 1 norm solution using linear programming (LP), we seek:…”

Noise-Resilient Single-Pixel Compressive Sensing with Single Photon Counting

“…These patterns under-sample the target's optical properties for image reconstruction, recognition, and assessment, albeit at the price of information loss. Recently, it has been shown that artificial intelligence (AI) and machine learning can be employed to process compressed data, where the quality of the reconstructed image is improved by deep learning and recurrent neural networks [28][29][30][31][32] .…”