Skip to main content

Research Repository

Advanced Search

Topological optimization of the DenseNet with pretrained-weights inheritance and genetic channel selection.

Fang, Zhenyu; Ren, Jinchang; Marshall, Stephen; Zhao, Huimin; Wang, Song; Li, Xuelong

Authors

Zhenyu Fang

Jinchang Ren

Stephen Marshall

Huimin Zhao

Song Wang

Xuelong Li



Abstract

Convolutional neural networks (CNNs) have been successfully applied in many computer vision applications, especially in image classification tasks, where most of the structures have been designed manually. With the aid of skip connection and dense connection, the depths of the models are becoming 'deeper' and the filters of layers are getting 'wider' in order to tackle the challenge of large-scale datasets. However, large-scale models in convolutional layers become inefficient due to the redundant channels from input feature maps. In this paper, we aim to automatically optimize the topology of the DenseNet, in which unnecessary convolutional kernels are reduced. To achieve this, we present a training pipeline that generates the network structure using a genetic algorithm. We first propose two encoding methods that can represent the structure of the model using a fixed-length binary string. A three-step based evolutionary process consisting of selection, crossover, and mutation is proposed to optimize the structure. We also present a pretrained weight inheritance method which can largely reduce the total time consumption of the genetic process. Experimental results have demonstrated that our proposed model can achieve comparable accuracy to the state-of-the-art models, across a wide range of image recognition and classification datasets, whilst significantly reducing the number of parameters.

Citation

FANG, Z, REN, J., MARSHALL, S., ZHAO, H., WANG, S. and LI, X. 2021. Topological optimization of the DenseNet with pretrained-weights inheritance and genetic channel selection. Pattern recognition [online], 109, article ID 107608. Available from: https://doi.org/10.1016/j.patcog.2020.107608

Journal Article Type Article
Acceptance Date Aug 19, 2020
Online Publication Date Aug 22, 2020
Publication Date Jan 31, 2021
Deposit Date Feb 22, 2021
Publicly Available Date Aug 23, 2021
Journal Pattern Recognition
Print ISSN 0031-3203
Publisher Elsevier
Peer Reviewed Peer Reviewed
Volume 109
Article Number 107608
DOI https://doi.org/10.1016/j.patcog.2020.107608
Keywords Deep convolutional neural networks; Genetic algorithms; Parameter reduction; Structure optimization; DenseNet
Public URL https://rgu-repository.worktribe.com/output/1084791

Files





You might also like



Downloadable Citations