Chair: Sarah Rajala, North Carolina State University (USA)
Thomas P. O'Rourke, University of Notre Dame (USA)
Robert L. Stevenson, University of Notre Dame (USA)
Yih-Fang Huang, University of Notre Dame (USA)
Lance C. Perez, University of Notre Dame (USA)
Daniel J. Costello Jr., University of Notre Dame (USA)
Many image communication systems have constraints on bandwidth, power and time which prohibit transmission of uncompressed raw image data. Compressed image formats, however, are extremely sensitive to bit errors which can seriously degrade the quality of the image at the receiver. A new list-based iterative trellis decoder is proposed which accepts feedback from a post-processor which can detect channel errors in the reconstructed image. Experimental results are shown which indicate the new decoder provides significant improvement over the standard Viterbi decoder.
William E. Lynch, Concordia University (CANADA)
Amy R. Reibman, AT&T Bell Laboratories
Bede Liu, Princeton University (USA)
Transform coding, a popular image compression strategy, results in two visible artifacts: blocking and ringing. These are both high frequency artifacts. Since images contain high frequency information the artifacts are removed using a space varying low pass filter as a post processor. Low frequency blocks and flat regions of blocks containng a strong edge are filtered. Low frequency blocks are identified in the transform coefficient domain; edge blocks are identified in the spatial domain. This does not require any alterations in the compressed bit stream. Improvement is demonstrated both subjectively and objectively.
Bin Zhu, University of Minnesota (USA)
Ahmed H. Tewfik, University of Minnesota (USA)
Omer Nezih Gerek, Bilkent University (TURKEY)
In this paper, we propose a novel approach for low bit rate perceptually transparent image compression. It exploits both frequency and spatial visual masking effects and uses a combination of Fourier and wavelet transforms to encode different bands. Frequency domain masking is computed by using a fine to coarse analysis step. Spatial domain masking is computed either by using Girod's model or a coarse to fine analysis step that accurately computes local contrast. A discrete cosine transform is used in conjunction with frequency domain masking to encode the low frequency bands. The medium and high frequency bands are encoded using spatial domain masking and a wavelet transform. The encoding of these bands is based on a recursive selection of the important edges in each band. It uses cross-band prediction to minimize bit rate. Experiments show the approach can achieve very high quality to nearly transparent compression at bit rates of 0.2 to 0.4 bits/pixel.
Matthew Crouse, University of Illinois (USA)
Kannan Ramchandran, University of Illinois (USA)
This paper introduces a novel, image-adaptive, encoding scheme for the baseline JPEG standard [wallace,jpegbook]. In particular, coefficient thresholding, JPEG quantization matrix (Q-matrix) optimization, and adaptive Huffman entropy-coding are jointly performed to maximize coded still-image quality within the constraints of the baseline JPEG syntax. Adaptive JPEG coding has been addressed in earlier works: in [kannan2], where fast rate-distortion (R-D) optimal coefficient thresholding was described, and in [gersho,hung], where R-D optimized Q-matrix selection was performed. By formulating an algorithm which optimizes these two operations jointly, we have obtained performance comparable to more complex, ``state-of-the-art'' coding schemes: for the ``Lenna'' image at 1 bpp, our algorithm has achieved a PSNR of 39.6 dB. This result represents a gain of 1.7 dB over JPEG with customized Huffman entropy coder, and even slightly exceeds the published performance of Shapiro's wavelet- based scheme [shapiro]. Furthermore, with the choice of appropriate visually-based error metrics, noticeable subjective improvement has been achieved as well.
Ping Wah Wong, Hewlett Packard Laboratories (USA)
An entropy constraint is introduced into a tree coding halftoner, so that one can control the degree of compressibility of the bi-level output images. The algorithm essentially trades image quality with compressibility as indicated by rate distortion theory. We demonstrate that this algorithm can generate halftone images that are of higher quality than error diffusion, and yet are also more amenable to compression than error diffused images.
Ephraim Feig, IBM Research
Heidi Peterson, IBM Research
Viresh Ratnakar, University of Wisconsin-Madison (USA)
This paper describes a new image compression technique, referred to as spatial predicition. Spatial prediction works in a manner similar to fractal-based image compression techniques, and is in fact a result of several experiments that we conducted to gain a better understanding of why fractal compression works. Spatial prediction compresses an image by storing, for each image block, either the quantized Discrete Cosine Transform (DCT) coefficients or the parameters of an affine transformation that constructs the block using another image block from the already encoded portion of the image. This technique does not require contractivity in the affine transformations and performs as well as or better than fractal compression. Spatial prediction does not out-perform pure DCT-based techniques (such as JPEG) in terms of PSNR/bit-rate tradeoff. However, at very low bit rates it results in far fewer blocky artifacts and markedly better visual quality.
M. Effros, California Institute of Technology
P.A. Chou, Xerox Palo Alto Research Center (USA)
We introduce a two-stage bit allocation algorithm analogous to the algorithm for weighted universal vector quantization (WUVQ) [Chou 1991],[Chou, Effros, and Gray 1994]. The encoder uses a collection of possible bit allocations (typically in the form of a collection of quantization matrices) rather than a single bit allocation (or single quantization matrix). We describe both an encoding algorithm for achieving optimal compression using a collection of bit allocations and a technique for designing locally optimal collections of bit allocations. We demonstrate performance on a JPEG style coder using the mean squared error (mse) distortion measure. On a sequence of medical brain scans, the algorithm achieves up to 2.5 dB improvement over a single bit allocation system, up to 5 dB improvement over a WUVQ with first- and second-stage vector dimensions equal to 16 and 4 respectively, and up to 12 dB improvement over an entropy constrained vector quantizer (ECVQ) using 4 dimensional vectors.
M. Das, Oakland University
D. L. Neuhoff, University of Michigan
C. L. Lin, Oakland University (USA)
This paper studies the characteristic properties of a specific class of near-lossless image compression schemes which consists of a lossless coder followed by a uniform scalar quantizer. Three specific instantiations of such schemes are investigated; namely, differential pulse code modulation, hierarchical interpolation, and two-dimensional space-varying multiplicative autoregressive coders. The compression gains attainable with such schemes are studied and results of experiments conducted on several medical images are presented.
S.J.P. Westen, Delft University of Technology (THE NETHERLANDS)
R.L. Lagendijk, Delft University of Technology (THE NETHERLANDS)
J. Biemond, Delft University of Technology (THE NETHERLANDS)
We propose a new measure of perceptual image quality based on a multiple channel human visual system (HVS) model for use in digital image compression. The model incorporates the HVS light sensitivity, spatial frequency and orientation sensitivity, and masking effects. The model is based on the concept of local band-limited contrast (LBC) in oriented spatial frequency bands. This concept leads to a simple masking function. The model has the flexibility to account for the changes in frequency sensitivity as a function of local luminance and is consistent with masking experiments using gratings and edges. Numerical scaling experiments with a test panel and a set a test images that were coded using different coding algorithms showed that the proposed measure correlates better with perceptual image quality than the conventional SNR measure.
Reiner Lenz, Linkoping University (SWEDEN)
Jonas Svanberg, Linkoping University (SWEDEN)
We use the representation theory of finite groups to simplify the Karhunen-Loeve Transform of systems with group theoretically defined symmetries. In this paper we focus on the dihedral group D(4) which is of importance for problems on square grids. For each group there is is a type of Fourier Transform which block- diagonalizes all operators that commute with the group operations. As a result all correlation matrices of processes with group theoretically defined symmetries are block-diagonalized. This simplifies the computation of the KLT considerably. For real world data the symmetry assumptions are never exactly fulfilled and the KLT based on the block-diagonal correlation matrix is only an approximation to the correct KLT. In the second part of the paper we compare several approximations to the KLT for a large data base consisting of blocks collected from a standard TV-channel. Finally we discuss some of the consequences for image coding applications.