A complexity comparison between multilayer perceptrons applied to on-sensor image compression

A multilayer perceptron (MLP) can be used to implement a vector quantizer (VQ) under severe constraints in the computational complexity allowed. Such constraints are typical in applications such as focal-plane image compression, in which we are interested in eliminating the analog-to-digital (A/D) c...

Full description

Saved in:
Bibliographic Details
Main Authors: Gomes, J.G.R.C., Mitra, S.K., de Figueiredo, R.J.P.
Format: Conference Proceeding
Language:English
Subjects:
Analog-digital conversion
Applied sciences
Artificial intelligence
Coding, codes
Computational complexity
Computer science
control theory
systems
Connectionism. Neural networks
Exact sciences and technology
Hardware
Image coding
Image converters
Image processing
Information, signal and communications theory
Kernel
Lattices
Multilayer perceptrons
Sampling, quantization
Signal and communications theory
Signal processing
Silicon
Streaming media
Telecommunications and information theory
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:A multilayer perceptron (MLP) can be used to implement a vector quantizer (VQ) under severe constraints in the computational complexity allowed. Such constraints are typical in applications such as focal-plane image compression, in which we are interested in eliminating the analog-to-digital (A/D) converters and mapping the analog data directly into a compressed bit stream, to save energy and silicon area. We compare a nonlinear MLP called the kernel lattice vector quantizer (KLVQ) and a clustering MLP known as the cluster-detection-and-labeling (CDL) network, with regard to their hardware requirements. We show that for similar rate-distortion performances, the KLVQ has complexity smaller than that of the CDL network.
ISSN:1520-6149
2379-190X
DOI:10.1109/ICASSP.2004.1327220