An image compression sensing method based on a sparse denoising self-coding network

Abstract

The invention claims an image compression sensing method based on a sparse denoising self-coding network, and belongs to the technical field of deep learning and image processing. The method comprisesthe following steps: 1, obtaining an original image signal x as training data, preprocessing the data and completing signal corrosion to obtain the formula as shown in the specification; 2, buildinga coding sub-network of a sparse denoising self-coding network, and obtaining a measurement value y by the image signal x through the coding sub-network; 3, setting up a decoding sub-network of the sparse denoising self-coding network, obtaining a reconstructed picture as shown in the specification by the measurement value y through the decoding sub-network, 4, introducing sparsity limitation, andgenerating a loss function JSDAE (W, b); and 5, carrying out joint training on the coding and decoding sub-networks through a back propagation algorithm, updating parameters and obtaining an optimalsparse denoising self-coding network. Sparsity limitation is added on the basis of the denoising self-coding network, image compression and reconstruction are integrated into a unified self-coding network framework, the quality of reconstructed images is effectively improved, and the reconstruction time is greatly shortened.

Description

technical field [0001] The invention belongs to the technical field of deep learning and image processing, and specifically relates to an image compression sensing method based on a sparse denoising self-encoding network, which effectively improves the quality of reconstructed images and greatly reduces reconstruction time. Background technique [0002] With the development of social informatization, the amount of data collected and processed has increased sharply, and the requirements for sensor sampling rate, storage device and transmission bandwidth are getting higher and higher. The traditional way of signal processing is to sample at a high rate and then store or transmit after compression, which will cause a lot of waste of sampled data. Thus the theory of compressed sensing emerged, which can collect signals at a rate much lower than the Nyquist sampling frequency, reconstruct the original signal with high precision, and complete compression while collecting the signa...

AI Technical Summary

Problems solved by technology

[0004] The invention aims to solve the problem that traditional compressed sensing image reconstruction takes a long time and the processing effect on noise pictures is poor

Images