A Novel Feature Fusion Method for Computing Image Aesthetic Quality
Computationally, the aesthetic quality of an image means that the model automatically scores the aesthetic level of the image. However, there are many factors that determine beauty or ugliness for photographic photos. Therefore, extracting a variety of representative aesthetic features and fusing th...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2020-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/9049106/ |
id |
doaj-ef27b0cc7b9f45169a11c1ab1c43aca9 |
---|---|
record_format |
Article |
spelling |
doaj-ef27b0cc7b9f45169a11c1ab1c43aca92021-03-30T01:36:02ZengIEEEIEEE Access2169-35362020-01-018630436305410.1109/ACCESS.2020.29837259049106A Novel Feature Fusion Method for Computing Image Aesthetic QualityXuewei Li0https://orcid.org/0000-0001-5336-7234Xueming Li1https://orcid.org/0000-0003-1058-2799Gang Zhang2https://orcid.org/0000-0002-1803-3180Xianlin Zhang3https://orcid.org/0000-0003-3905-2062Beijing Key Laboratory of Network System and Network Culture, Beijing University of Posts and Telecommunications, Beijing, ChinaBeijing Key Laboratory of Network System and Network Culture, Beijing University of Posts and Telecommunications, Beijing, ChinaSchool of Armament Science and Technology, Space Engineering University, Beijing, ChinaBeijing Key Laboratory of Network System and Network Culture, Beijing University of Posts and Telecommunications, Beijing, ChinaComputationally, the aesthetic quality of an image means that the model automatically scores the aesthetic level of the image. However, there are many factors that determine beauty or ugliness for photographic photos. Therefore, extracting a variety of representative aesthetic features and fusing these features are still difficult tasks. In this paper, we design a two-stream network to calculate the aesthetic quality of the image. The upper stream of the network is an improved network with the SEResNet-50 and six skip connections added, which can improve the performance of the model without training to obtain deep convolutional neural network features. The lower stream of the network consists of the proposed algorithms for handcrafted extracting aesthetic features and multiple convolution layers to extract the aesthetic features. Finally, to fuse the features of the two-stream network without adding feature dimensions, a novel feature fusion layer is proposed. The results show that this novel feature fusion method can calculate results close to the artificial aesthetic evaluation.https://ieeexplore.ieee.org/document/9049106/Deep convolutional neural networksfeature fusionhandcrafted aesthetic featuresimage aesthetics quality assessment |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Xuewei Li Xueming Li Gang Zhang Xianlin Zhang |
spellingShingle |
Xuewei Li Xueming Li Gang Zhang Xianlin Zhang A Novel Feature Fusion Method for Computing Image Aesthetic Quality IEEE Access Deep convolutional neural networks feature fusion handcrafted aesthetic features image aesthetics quality assessment |
author_facet |
Xuewei Li Xueming Li Gang Zhang Xianlin Zhang |
author_sort |
Xuewei Li |
title |
A Novel Feature Fusion Method for Computing Image Aesthetic Quality |
title_short |
A Novel Feature Fusion Method for Computing Image Aesthetic Quality |
title_full |
A Novel Feature Fusion Method for Computing Image Aesthetic Quality |
title_fullStr |
A Novel Feature Fusion Method for Computing Image Aesthetic Quality |
title_full_unstemmed |
A Novel Feature Fusion Method for Computing Image Aesthetic Quality |
title_sort |
novel feature fusion method for computing image aesthetic quality |
publisher |
IEEE |
series |
IEEE Access |
issn |
2169-3536 |
publishDate |
2020-01-01 |
description |
Computationally, the aesthetic quality of an image means that the model automatically scores the aesthetic level of the image. However, there are many factors that determine beauty or ugliness for photographic photos. Therefore, extracting a variety of representative aesthetic features and fusing these features are still difficult tasks. In this paper, we design a two-stream network to calculate the aesthetic quality of the image. The upper stream of the network is an improved network with the SEResNet-50 and six skip connections added, which can improve the performance of the model without training to obtain deep convolutional neural network features. The lower stream of the network consists of the proposed algorithms for handcrafted extracting aesthetic features and multiple convolution layers to extract the aesthetic features. Finally, to fuse the features of the two-stream network without adding feature dimensions, a novel feature fusion layer is proposed. The results show that this novel feature fusion method can calculate results close to the artificial aesthetic evaluation. |
topic |
Deep convolutional neural networks feature fusion handcrafted aesthetic features image aesthetics quality assessment |
url |
https://ieeexplore.ieee.org/document/9049106/ |
work_keys_str_mv |
AT xueweili anovelfeaturefusionmethodforcomputingimageaestheticquality AT xuemingli anovelfeaturefusionmethodforcomputingimageaestheticquality AT gangzhang anovelfeaturefusionmethodforcomputingimageaestheticquality AT xianlinzhang anovelfeaturefusionmethodforcomputingimageaestheticquality AT xueweili novelfeaturefusionmethodforcomputingimageaestheticquality AT xuemingli novelfeaturefusionmethodforcomputingimageaestheticquality AT gangzhang novelfeaturefusionmethodforcomputingimageaestheticquality AT xianlinzhang novelfeaturefusionmethodforcomputingimageaestheticquality |
_version_ |
1724186848259997696 |