High-order image subsampling using feedforward artificial neural networks

Author

Dumitras, Adriana ; Kossentini, Faouzi

Author_Institution

AT&T Labs.-Res., Middletown, NJ, USA

Volume

10

Issue

3

fYear

2001

fDate

3/1/2001 12:00:00 AM

Firstpage

427

Lastpage

435

Abstract

We propose a method for high-order image subsampling using feedforward artificial neural networks (FANNs). In our method, the high-order subsampling process is decomposed into a sequence of first-order subsampling stages. The first stage employs a tridiagonally symmetrical FANN, which is obtained by applying the design algorithm introduced by Dumitras and Kossentini (see IEEE Trans. Signal Processing, vol.48, p.1446-55, 2000). The second stage employs a small fully connected FANN. The algorithm used to train both FANNs employs information about local edges (extracted using pattern matching) to perform effective subsampling of both high detail and smooth image areas. We show that our multistage first-order subsampling method achieves excellent speed-performance tradeoffs, and it consistently outperforms traditional lowpass filtering and subsampling methods both subjectively and objectively

Keywords

feedforward neural nets; image sampling; learning (artificial intelligence); pattern matching; ANN; design algorithm; feedforward artificial neural networks; first-order subsampling stages; fully connected FANN; high detail image areas; high-order image subsampling; local edges; multistage first-order subsampling; neural network training; pattern matching; smooth image areas; speed-performance tradeoffs; tridiagonally symmetrical FANN; Algorithm design and analysis; Artificial neural networks; Data mining; Filtering; Image converters; Image reconstruction; Image sampling; Neural networks; Pattern matching; Signal sampling;

fLanguage

English

Journal_Title

Image Processing, IEEE Transactions on

Publisher

ieee

ISSN

1057-7149

Type

jour

DOI

10.1109/83.908518

Filename

908518