A Fully Embedded Two-Stage Coder for Hyperspectral Near-Lossless Compression

Abstract

Abstract-This letter proposes a near-lossless coder for hyperspectral images. The coding technique is fully embedded and minimizes the distortion in the l2 norm initially and in the l∞ norm subsequently. Based on a two-stage near-lossless compression scheme, it includes a lossy and a near-lossless layer. The novelties are: the observation of the convergence of the entropy of the residuals in the original domain and in the spectral-spatial transformed domain; and an embedded near-lossless layer. These contributions enable a progressive transmission while optimising both SNR and PAE performance. The embeddedness is accomplished by bitplane encoding plus arithmetic encoding. Experimental results suggest that the proposed method yields a highly competitive coding performance for hyperspectral images, outperforming multi-component JPEG2000 for l∞ norm and pairing its performance for l2 norm, and also outperforming M-CALIC in the near-lossless case -for PAE ≥ 5-.

Regression Wavelet Analysis for Near-Lossless Remote Sensing Data Compression

“…Concerning the bitrate performance, a cursory glance to Table II shows the good behaviour of [23] (with the exception of results for the CH corpus in the lossless regime). However, [23] applies a Karhunen-Loève-Transform (KLT) before JPEG 2000 lossy layer. KLT efficiently decorrelates the spectral dimension of a scene, but entails a high computational complexity and a non-negligible side information.…”

“…Table II reports bitrates and quality (SNR) results for the 9 scenes presented in the original work of Beerten et al [23] in comparison with our approach. UA12 refers to the UA sensor that has stored the scenes with a bit-depth of 12 bpppc.…”

“…Both [23] and our approach use a bitplane-based entropy coder that allows for some degree of embeddedness and enables progressive lossy-to-lossless/near-lossless transmission. Concerning the quality of the reconstructed scenes, given any error tolerance value Λ, our proposal is always superior, as happened for the case of comparing against prediction-based coding techniques.…”

“…Methods such as [19] and [20] do not include any selection criterion to determine the lossy bitrate. Other methods such as [21], [22] and [23] achieve competitive coding performance after using computationally expensive iterative approaches to identify the optimal lossy bitrate. The embedded two-stage near-lossless coder [23] yields the stateof-the-art compression performance in this category.…”

A survey and evaluation of bior wavelet based compression techniques

“…Multiresolution graph Fourier transform for compression of piecewise smooth images; in this research the authors focused on graph Fourier transforms (GFTs) to minimize the total signal representation cost of each pixel (Wei et al, 2015). A fully embedded two-stage coder for hyperspectral near-lossless compression; this research exploited a two stage near-lossless coder for hyperspectral images (Beerten et al, 2015). Learning-based joint super-resolution and deblocking for a highly compressed image; in this paper, the author investigates how to preserve artifacts using the joint super-resolution technique for highly compressed images (Kang et al, 2015).…”