A Simple and Effective Convolutional Filter Pruning based on Filter Dissimilarity Analysis

F. Erick, Shrutika Sawant, Stephan Göb, N. Holzer, E. Lang, Th. Götz, Th. Götz, Th. Götz

2022

Abstract

In this paper, a simple and effective filter pruning method is proposed to simplify the deep convolutional neural network (CNN) and accelerate learning. The proposed method selects the important filters and discards the unimportant ones based on filter dissimilarity analysis. The proposed method searches for filters with decent representative ability and less redundancy, discarding the others. The representative ability and redundancy contained in the filter is evaluated by its correlation with currently selected filters and left over unselected filters. Moreover, the proposed method uses an iterative procedure, so that less representative filters can be discarded evenly from the entire model. The experimental analysis confirmed that a simple filter pruning method can reduce floating point operations (FLOPs) of TernausNet by up to 89.65% on an INRIA Aerial Image Labeling dataset with an only marginal drop in the original accuracy. Furthermore, the proposed method shows promising results in comparison with other state-of-the-art methods.

Paper Citation

in Harvard Style

Erick F., Sawant S., Göb S., Holzer N., Lang E. and Götz T. (2022). A Simple and Effective Convolutional Filter Pruning based on Filter Dissimilarity Analysis. In Proceedings of the 14th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART, ISBN 978-989-758-547-0, pages 139-145. DOI: 10.5220/0010786400003116

in Bibtex Style

@conference{icaart22,
author={F. Erick and Shrutika Sawant and Stephan Göb and N. Holzer and E. Lang and Th. Götz},
title={A Simple and Effective Convolutional Filter Pruning based on Filter Dissimilarity Analysis},
booktitle={Proceedings of the 14th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART,},
year={2022},
pages={139-145},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010786400003116},
isbn={978-989-758-547-0},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 14th International Conference on Agents and Artificial Intelligence - Volume 3: ICAART,
TI - A Simple and Effective Convolutional Filter Pruning based on Filter Dissimilarity Analysis
SN - 978-989-758-547-0
AU - Erick F.
AU - Sawant S.
AU - Göb S.
AU - Holzer N.
AU - Lang E.
AU - Götz T.
PY - 2022
SP - 139
EP - 145
DO - 10.5220/0010786400003116