Artificial Color Constancy via GoogLeNet with Angular Loss Function
Color constancy is the ability of the human visual system to perceive colors unchanged independently of illumination. Giving a machine this feature will be beneficial in many fields where chromatic information is used. Particularly, it significantly improves scene understanding and object recognitio...
Main Author: | |
---|---|
Format: | Article |
Language: | English |
Published: |
Taylor & Francis Group
2020-07-01
|
Series: | Applied Artificial Intelligence |
Online Access: | http://dx.doi.org/10.1080/08839514.2020.1730630 |
_version_ | 1797684897815461888 |
---|---|
author | Oleksii Sidorov |
author_facet | Oleksii Sidorov |
author_sort | Oleksii Sidorov |
collection | DOAJ |
description | Color constancy is the ability of the human visual system to perceive colors unchanged independently of illumination. Giving a machine this feature will be beneficial in many fields where chromatic information is used. Particularly, it significantly improves scene understanding and object recognition.In this article, we propose a transfer learning-based algorithm, which has two main features: accuracy higher than many state-of-the-art algorithms and simplicity of implementation. Despite the fact that GoogLeNet was used in the experiments, the given approach may be applied to any convolutional neural networks. Additionally, we discuss the design of a new loss function oriented specifically to this problem and propose a few of the most suitable options. |
first_indexed | 2024-03-12T00:36:27Z |
format | Article |
id | doaj.art-a8bc0b46f85e426a91de2201a005707d |
institution | Directory Open Access Journal |
issn | 0883-9514 1087-6545 |
language | English |
last_indexed | 2024-03-12T00:36:27Z |
publishDate | 2020-07-01 |
publisher | Taylor & Francis Group |
record_format | Article |
series | Applied Artificial Intelligence |
spelling | doaj.art-a8bc0b46f85e426a91de2201a005707d2023-09-15T09:33:58ZengTaylor & Francis GroupApplied Artificial Intelligence0883-95141087-65452020-07-0134964365510.1080/08839514.2020.17306301730630Artificial Color Constancy via GoogLeNet with Angular Loss FunctionOleksii Sidorov0NTNUColor constancy is the ability of the human visual system to perceive colors unchanged independently of illumination. Giving a machine this feature will be beneficial in many fields where chromatic information is used. Particularly, it significantly improves scene understanding and object recognition.In this article, we propose a transfer learning-based algorithm, which has two main features: accuracy higher than many state-of-the-art algorithms and simplicity of implementation. Despite the fact that GoogLeNet was used in the experiments, the given approach may be applied to any convolutional neural networks. Additionally, we discuss the design of a new loss function oriented specifically to this problem and propose a few of the most suitable options.http://dx.doi.org/10.1080/08839514.2020.1730630 |
spellingShingle | Oleksii Sidorov Artificial Color Constancy via GoogLeNet with Angular Loss Function Applied Artificial Intelligence |
title | Artificial Color Constancy via GoogLeNet with Angular Loss Function |
title_full | Artificial Color Constancy via GoogLeNet with Angular Loss Function |
title_fullStr | Artificial Color Constancy via GoogLeNet with Angular Loss Function |
title_full_unstemmed | Artificial Color Constancy via GoogLeNet with Angular Loss Function |
title_short | Artificial Color Constancy via GoogLeNet with Angular Loss Function |
title_sort | artificial color constancy via googlenet with angular loss function |
url | http://dx.doi.org/10.1080/08839514.2020.1730630 |
work_keys_str_mv | AT oleksiisidorov artificialcolorconstancyviagooglenetwithangularlossfunction |