Abstract:In many professional fields, such as medicine, remote sensing and sciences, users often demand image compression methods to be mathematically lossless. But lossless image coding has a rather low compression ratio (around 2: 1 for natural images). The only known technique to achieve significant compression while meeting the stringent fidelity requirements is the methodology of -constrained coding that was developed and standardized in nineties. We make a major progress in -constrained image coding after two decades, by developing a novel CNN-based soft -constrained decoding method. The new method repairs compression defects by using a restoration CNN called the to map a conventionally decoded image to the latent image. A unique strength of the is its ability to enforce a tight error bound on a per pixel basis. As such, no small distinctive structures of the original image can be dropped or distorted, even if they are statistical outliers that are otherwise sacrificed by mainstream CNN restoration methods. More importantly, this research ushers in a new image compression system of -constrained encoding and deep soft decoding (). The approach beats the best of existing lossy image compression methods (eg, BPG, WebP, etc.) not only in but also in error metric and perceptual quality, for bit rates near the threshold of perceptually transparent reconstruction. Operationally, the new compression system is practical, with a low-complexity real-time encoder and a cascade decoder consisting of a fast initial decoder and an optional CNN soft decoder.

Decoding of Compressive Data Pages for Optical Data Storage Utilizing FFDNet

Coded Aperture-Based Compressive Data Page for Optical Data Storage.

Foveated JND Model Based on Stereo Vision and Its Application in Image Compression with Hardware Implementation

Data Compression and Storage under High Speed Network

Soft Decoding of Light Field Images Using Pocs and Fast Graph Spectrayl Filters.

High-accuracy data readout in multi-dimensional optical data storage using convolutional neural networks

High noise margin decoding of holographic data page based on compressed sensing

An Encoded Photodetector Array for High-frame-rate Compressed Sensing and Computational Imaging

Computationally-Efficient Neural Image Compression with Shallow Decoders

Decoding Method in the Amplitude-Modulated Collinear Holographic Data Storage System

Integrated photonic encoder for low power and high-speed image processing

Fast Algorithm for DEM Lossless Compression

Low-cost and Fast Partial Decoding of Compressed Videos.

Ultra High Fidelity Image Compression with -constrained Encoding and Deep Decoding

DSCR-Net: A Diffractive Sensing and Complex-Valued Reconstruction Network for Compressive Sensing.

Scientific data lossless compression using fast neural network

Gaze-contingent Efficient Hologram Compression for Foveated Near-Eye Holographic Displays.

Coded Aperture Compression Temporal Imaging Based on a Dual-Mask and Deep Denoiser

Content Adaptive Latents and Decoder for Neural Image Compression

Data Compression of Light Field Using Multiscale Edges

A comparative study for massive data compression in long-distance distributed optical fiber sensing systems