Resfeats: Residual network based features for image classification

© 2017 IEEE. Deep residual networks have recently emerged as the state-of-the-art architecture in image classification and object detection. In this paper, we propose new image features (called ResFeats) extracted from the last convolutional layer of the deep residual networks pre-trained on ImageNe...

Full description

Bibliographic Details
Main Authors: Mahmood, A., Bennamoun, M., An, Senjian, Sohel, F.
Format: Conference Paper
Published: 2018
Online Access:http://hdl.handle.net/20.500.11937/70045
_version_ 1848762200888442880
author Mahmood, A.
Bennamoun, M.
An, Senjian
Sohel, F.
author_facet Mahmood, A.
Bennamoun, M.
An, Senjian
Sohel, F.
author_sort Mahmood, A.
building Curtin Institutional Repository
collection Online Access
description © 2017 IEEE. Deep residual networks have recently emerged as the state-of-the-art architecture in image classification and object detection. In this paper, we propose new image features (called ResFeats) extracted from the last convolutional layer of the deep residual networks pre-trained on ImageNet. We propose to use ResFeats for diverse image classification tasks namely, object classification, scene classification and coral classification and show that ResFeats consistently perform better than their CNN counterparts on these classification tasks. Since the ResFeats are large feature vectors, we explore dimensionality reduction methods. Experimental results are provided to show the effectiveness of ResFeats with state-of-the-art classification accuracies on Caltech-101, Caltech-256 and MLC datasets and a significant performance improvement on MIT-67 dataset compared to the widely used CNN features.
first_indexed 2025-11-14T10:43:48Z
format Conference Paper
id curtin-20.500.11937-70045
institution Curtin University Malaysia
institution_category Local University
last_indexed 2025-11-14T10:43:48Z
publishDate 2018
recordtype eprints
repository_type Digital Repository
spelling curtin-20.500.11937-700452018-08-08T04:57:01Z Resfeats: Residual network based features for image classification Mahmood, A. Bennamoun, M. An, Senjian Sohel, F. © 2017 IEEE. Deep residual networks have recently emerged as the state-of-the-art architecture in image classification and object detection. In this paper, we propose new image features (called ResFeats) extracted from the last convolutional layer of the deep residual networks pre-trained on ImageNet. We propose to use ResFeats for diverse image classification tasks namely, object classification, scene classification and coral classification and show that ResFeats consistently perform better than their CNN counterparts on these classification tasks. Since the ResFeats are large feature vectors, we explore dimensionality reduction methods. Experimental results are provided to show the effectiveness of ResFeats with state-of-the-art classification accuracies on Caltech-101, Caltech-256 and MLC datasets and a significant performance improvement on MIT-67 dataset compared to the widely used CNN features. 2018 Conference Paper http://hdl.handle.net/20.500.11937/70045 10.1109/ICIP.2017.8296551 restricted
spellingShingle Mahmood, A.
Bennamoun, M.
An, Senjian
Sohel, F.
Resfeats: Residual network based features for image classification
title Resfeats: Residual network based features for image classification
title_full Resfeats: Residual network based features for image classification
title_fullStr Resfeats: Residual network based features for image classification
title_full_unstemmed Resfeats: Residual network based features for image classification
title_short Resfeats: Residual network based features for image classification
title_sort resfeats: residual network based features for image classification
url http://hdl.handle.net/20.500.11937/70045