Neural network approach to DPCM system design for image coding

Manikopoulos, C.N.

doi:10.1049/ip-i-2.1992.0067

Cited by 31 publications

(8 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Predictive performance with neural networks is claimed to outperform the conventional optimum linear predictors by about 4.17 and 3.74 dB for two test images [44]. Further research, especially for non-linear networks, is encouraged by the reported results to optimize their learning rules for prediction of those images whose contents are subject to abrupt statistical changes.…”

Section: A Predictive Coding Neural Networkmentioning

confidence: 97%

“…Non-linear predictive coding, however, is very limited due to the difficulties involved in optimizing the coefficients extraction to obtain the best possible predictive values. Under this circumstance, a neural network provides a very promising approach in optimizing non-linear predictive coding [43,44]. Based on a linear AR model, a multilayer perceptron neural network can be constructed to achieve the design of its corresponding non-linear predictor as shown in Fig.…”

Section: A Predictive Coding Neural Networkmentioning

confidence: 99%

See 1 more Smart Citation

Hybrid Approaches to Image Coding: A Review

Rehna¹,

Jeyakumar²

2011

IJACSA

View full text Add to dashboard Cite

Abstract-Now a days, the digital world is most focused on storage space and speed. With the growing demand for better bandwidth utilization, efficient image data compression techniques have emerged as an important factor for image data transmission and storage. To date, different approaches to image compression have been developed like the classical predictive coding, popular transform coding and vector quantization. Several second generation coding schemes or the segmentation based schemes are also gaining popularity. Practically efficient compression systems based on hybrid coding which combines the advantages of different traditional methods of image coding have also being developed over the years. In this paper, different hybrid approaches to image compression are discussed. Hybrid coding of images, in this context, deals with combining two or more traditional approaches to enhance the individual methods and achieve better quality reconstructed images with higher compression ratio. Literature on hybrid techniques of image coding over the past years is also reviewed. An attempt is made to highlight the neuro-wavelet approach for enhancing coding efficiency.

show abstract

Section: A Predictive Coding Neural Networkmentioning

confidence: 97%

Section: A Predictive Coding Neural Networkmentioning

confidence: 99%

Hybrid Approaches to Image Coding: A Review

Rehna¹,

Jeyakumar²

2011

IJACSA

View full text Add to dashboard Cite

show abstract

“…Unfortunately, there are not as mathematically tractable as linear predictors. In addition, they are time consuming and usually impossible to design optimal nonlinear predictors [34].…”

Section: One Dimensional Two Dimensional Casualmentioning

confidence: 99%

Image compression techniques: A survey in lossless and lossy algorithms

2018

View full text Add to dashboard Cite

The bandwidth of the communication networks has been increased continuously as results of technological advances. However, the introduction of new services and the expansion of the existing ones have resulted in even higher demand for the bandwidth. This explains the many efforts currently being invested in the area of data compression. The primary goal of these works is to develop techniques of coding information sources such as speech, image and video to reduce the number of bits required to represent a source without significantly degrading its quality. With the large increase in the generation of digital image data, there has been a correspondingly large increase in research activity in the field of image compression. The goal is to represent an image in the fewest number of bits without losing the essential information content within. Images carry three main type of information: redundant, irrelevant, and useful. Redundant information is the deterministic part of the information, which can be reproduced without loss from other information contained in the image. Irrelevant information is the part of information that has enormous details, which are beyond the limit of perceptual significance (i.e., psychovisual redundancy). Useful information, on the other hand, is the part of information, which is neither redundant nor irrelevant. Human usually observes decompressed images. Therefore, their fidelities are subject to the capabilities and limitations of the Human Visual System.

show abstract

“…Another approach is to take advantage of the gradient descent properties of the backpropagation algorithm for the calculation of the optimal predictor for a specific nonlinear model. Manikopoulos [22], [23] has used a nonlinear predictor based on the discrete-time Volterra expansion of a generalized nonlinear predictor model. The justification for using a nonlinear approach is that linear AR image models do not adequately account for sharply defined structures such as edges in images.…”

Section: B Higher-order Predictorsmentioning

confidence: 99%

Neural network approaches to image compression

1995

View full text Add to dashboard Cite

This paper presents a tutorial overview of neural networks as signal processing tools for image compression. They are well suited to the problem of image compression due to their massively parallel and distributed architecture. Their characteristics are analogous to some of the features of our own visual system, which allow us to process visual information with much ease. For example, multilayer perceptrons can be used as nonlinear predictors in differential pulse-code modulation (DPCM). Such predictors have been shown to increase the predictive gain relative to a linear predictor. Another active area of research is in the application of Hebbian learning to the extraction of principal components, which are the basis vectors for the optimal linear Karhunen-Loève transform (KLT). These learning algorithms are iterative, have some computational advantages over standard eigendecomposition techniques, and can be made to adapt to changes in the input signal. Yet another model, the self-organizing feature map (SOFM), has been used with a great deal of success in the design of codebooks for vector quantization (VQ). The resulting codebooks are less sensitive to initial conditions than the standard LBG algorithm, and the topological ordering of the entries can be exploited to further increase coding efficiency and reduce computational complexity.

show abstract

Neural network approach to DPCM system design for image coding

Cited by 31 publications

References 2 publications

Hybrid Approaches to Image Coding: A Review

Hybrid Approaches to Image Coding: A Review

Image compression techniques: A survey in lossless and lossy algorithms

Neural network approaches to image compression

Contact Info

Product

Resources

About