Accelerating Image Classification using Feature Map Similarity in Convolutional Neural Networks
Convolutional neural networks (CNNs) have greatly improved image classification performance. However, the extensive time required for classification owing to the large amount of computation involved, makes it unsuitable for application to low-performance devices. To speed up image classification, we...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2018-12-01
|
Series: | Applied Sciences |
Subjects: | |
Online Access: | http://www.mdpi.com/2076-3417/9/1/108 |
id |
doaj-62edfe56d7c64fa98568549b65ad32ba |
---|---|
record_format |
Article |
spelling |
doaj-62edfe56d7c64fa98568549b65ad32ba2020-11-24T20:46:28ZengMDPI AGApplied Sciences2076-34172018-12-019110810.3390/app9010108app9010108Accelerating Image Classification using Feature Map Similarity in Convolutional Neural NetworksKeunyoung Park0Doo-Hyun Kim1Department of Software, Konkuk University, Seoul 05029, KoreaDepartment of Software, Konkuk University, Seoul 05029, KoreaConvolutional neural networks (CNNs) have greatly improved image classification performance. However, the extensive time required for classification owing to the large amount of computation involved, makes it unsuitable for application to low-performance devices. To speed up image classification, we propose a cached CNN, which can classify input images based on similarity with previously input images. Because the feature maps extracted from the CNN kernel represent the intensity of features, images with a similar intensity can be classified into the same class. In this study, we cache class labels and feature vectors extracted from feature maps for images classified by the CNN. Then, when a new image is input, its class label is output based on its similarity with the cached feature vectors. This process can be performed at each layer; hence, if the classification is successful, there is no need to perform the remaining convolution layer operations. This reduces the required classification time. We performed experiments to measure and evaluate the cache hit rate, precision, and classification time.http://www.mdpi.com/2076-3417/9/1/108image classificationconvolutional neural networkfeature mapcosine similarity |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Keunyoung Park Doo-Hyun Kim |
spellingShingle |
Keunyoung Park Doo-Hyun Kim Accelerating Image Classification using Feature Map Similarity in Convolutional Neural Networks Applied Sciences image classification convolutional neural network feature map cosine similarity |
author_facet |
Keunyoung Park Doo-Hyun Kim |
author_sort |
Keunyoung Park |
title |
Accelerating Image Classification using Feature Map Similarity in Convolutional Neural Networks |
title_short |
Accelerating Image Classification using Feature Map Similarity in Convolutional Neural Networks |
title_full |
Accelerating Image Classification using Feature Map Similarity in Convolutional Neural Networks |
title_fullStr |
Accelerating Image Classification using Feature Map Similarity in Convolutional Neural Networks |
title_full_unstemmed |
Accelerating Image Classification using Feature Map Similarity in Convolutional Neural Networks |
title_sort |
accelerating image classification using feature map similarity in convolutional neural networks |
publisher |
MDPI AG |
series |
Applied Sciences |
issn |
2076-3417 |
publishDate |
2018-12-01 |
description |
Convolutional neural networks (CNNs) have greatly improved image classification performance. However, the extensive time required for classification owing to the large amount of computation involved, makes it unsuitable for application to low-performance devices. To speed up image classification, we propose a cached CNN, which can classify input images based on similarity with previously input images. Because the feature maps extracted from the CNN kernel represent the intensity of features, images with a similar intensity can be classified into the same class. In this study, we cache class labels and feature vectors extracted from feature maps for images classified by the CNN. Then, when a new image is input, its class label is output based on its similarity with the cached feature vectors. This process can be performed at each layer; hence, if the classification is successful, there is no need to perform the remaining convolution layer operations. This reduces the required classification time. We performed experiments to measure and evaluate the cache hit rate, precision, and classification time. |
topic |
image classification convolutional neural network feature map cosine similarity |
url |
http://www.mdpi.com/2076-3417/9/1/108 |
work_keys_str_mv |
AT keunyoungpark acceleratingimageclassificationusingfeaturemapsimilarityinconvolutionalneuralnetworks AT doohyunkim acceleratingimageclassificationusingfeaturemapsimilarityinconvolutionalneuralnetworks |
_version_ |
1716812490342400000 |