Abstract
Affective image classification has attracted much attention in recent years. However, the production of more exact classifiers depends on the quality of the sample database. In this study, we analyzed various existing databases used for affective image classification and we tried to improve the quality of the learning data by combining existing databases in several different ways. We found that existing image databases cannot cover the overall range of the arousal-valence plane. Thus, to obtain a wider distribution of emotion labels from images, we conducted a crowd-sourcing-based user study with Amazon Mechanical Turk. We aimed to construct several different versions of affective image classifiers by using different combinations of existing databases, instead of using one. We used low-level features in our classification experiments to explore the discriminatory properties of emotion categories. We report the results of intermediate comparisons using different combinations of databases to evaluate the performance of this approach.
| Original language | English |
|---|---|
| Title of host publication | ACHI 2015 - 8th International Conference on Advances in Computer-Human Interactions |
| Editors | Leslie Miller, Alma Leora Culen |
| Publisher | International Academy, Research and Industry Association, IARIA |
| Pages | 211-212 |
| Number of pages | 2 |
| ISBN (Electronic) | 9781612083827 |
| Publication status | Published - 2015 Jan 1 |
| Event | 8th International Conference on Advances in Computer-Human Interactions, ACHI 2015 - Lisbon, Portugal Duration: 2015 Feb 22 → 2015 Feb 27 |
Publication series
| Name | ACHI 2015 - 8th International Conference on Advances in Computer-Human Interactions |
|---|
Other
| Other | 8th International Conference on Advances in Computer-Human Interactions, ACHI 2015 |
|---|---|
| Country | Portugal |
| City | Lisbon |
| Period | 15/2/22 → 15/2/27 |
Fingerprint
All Science Journal Classification (ASJC) codes
- Human-Computer Interaction
Cite this
}
Combining image databases for affective image classification. / Kim, Hye Rin; Lee, In Kwon.
ACHI 2015 - 8th International Conference on Advances in Computer-Human Interactions. ed. / Leslie Miller; Alma Leora Culen. International Academy, Research and Industry Association, IARIA, 2015. p. 211-212 (ACHI 2015 - 8th International Conference on Advances in Computer-Human Interactions).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
TY - GEN
T1 - Combining image databases for affective image classification
AU - Kim, Hye Rin
AU - Lee, In Kwon
PY - 2015/1/1
Y1 - 2015/1/1
N2 - Affective image classification has attracted much attention in recent years. However, the production of more exact classifiers depends on the quality of the sample database. In this study, we analyzed various existing databases used for affective image classification and we tried to improve the quality of the learning data by combining existing databases in several different ways. We found that existing image databases cannot cover the overall range of the arousal-valence plane. Thus, to obtain a wider distribution of emotion labels from images, we conducted a crowd-sourcing-based user study with Amazon Mechanical Turk. We aimed to construct several different versions of affective image classifiers by using different combinations of existing databases, instead of using one. We used low-level features in our classification experiments to explore the discriminatory properties of emotion categories. We report the results of intermediate comparisons using different combinations of databases to evaluate the performance of this approach.
AB - Affective image classification has attracted much attention in recent years. However, the production of more exact classifiers depends on the quality of the sample database. In this study, we analyzed various existing databases used for affective image classification and we tried to improve the quality of the learning data by combining existing databases in several different ways. We found that existing image databases cannot cover the overall range of the arousal-valence plane. Thus, to obtain a wider distribution of emotion labels from images, we conducted a crowd-sourcing-based user study with Amazon Mechanical Turk. We aimed to construct several different versions of affective image classifiers by using different combinations of existing databases, instead of using one. We used low-level features in our classification experiments to explore the discriminatory properties of emotion categories. We report the results of intermediate comparisons using different combinations of databases to evaluate the performance of this approach.
UR - http://www.scopus.com/inward/record.url?scp=84966550427&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84966550427&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84966550427
T3 - ACHI 2015 - 8th International Conference on Advances in Computer-Human Interactions
SP - 211
EP - 212
BT - ACHI 2015 - 8th International Conference on Advances in Computer-Human Interactions
A2 - Miller, Leslie
A2 - Culen, Alma Leora
PB - International Academy, Research and Industry Association, IARIA
ER -