Abstract
Bag-of-words (BoW) is one of the most successful methods for object categorization. This paper proposes a statistical codeword selection algorithm where the best subset is selected from the initial codewords based on the statistical characteristics of codewords. For this purpose, we defined two types of codeword-confidences: cross- and within-category confidences. The cross- and within-category confidences eliminate indistinctive codewords across categories and inconsistent codewords within each category, respectively. An informative subset of codewords is then selected based on these two codeword-confidences. The experimental evaluation for a scene categorization dataset and a Caltech-101 dataset shows that the proposed method improves the categorization performance up to 10% in terms of error rate reduction when cooperated with BoW, sparse coding (SC), and locality-constrained liner coding (LLC). Furthermore, the codeword size is reduced by 50% leading a low computational complexity.
| Original language | English |
|---|---|
| Title of host publication | Proceedings of SPIE-IS and T Electronic Imaging - Intelligent Robots and Computer Vision XXXII |
| Subtitle of host publication | Algorithms and Techniques |
| Editors | Juha Roning, David Casasent |
| Publisher | SPIE |
| ISBN (Electronic) | 9781628414967 |
| DOIs | |
| Publication status | Published - 2015 Jan 1 |
| Event | Intelligent Robots and Computer Vision XXXII: Algorithms and Techniques - San Francisco, United States Duration: 2015 Feb 9 → 2015 Feb 10 |
Publication series
| Name | Proceedings of SPIE - The International Society for Optical Engineering |
|---|---|
| Volume | 9406 |
| ISSN (Print) | 0277-786X |
| ISSN (Electronic) | 1996-756X |
Other
| Other | Intelligent Robots and Computer Vision XXXII: Algorithms and Techniques |
|---|---|
| Country | United States |
| City | San Francisco |
| Period | 15/2/9 → 15/2/10 |
Fingerprint
All Science Journal Classification (ASJC) codes
- Electronic, Optical and Magnetic Materials
- Condensed Matter Physics
- Computer Science Applications
- Applied Mathematics
- Electrical and Electronic Engineering
Cite this
}
Statistical approach for supervised codeword selection. / Park, Kihong; Ryu, Seungchul; Kim, Seungryong; Sohn, Kwanghoon.
Proceedings of SPIE-IS and T Electronic Imaging - Intelligent Robots and Computer Vision XXXII: Algorithms and Techniques. ed. / Juha Roning; David Casasent. SPIE, 2015. 940609 (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 9406).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
TY - GEN
T1 - Statistical approach for supervised codeword selection
AU - Park, Kihong
AU - Ryu, Seungchul
AU - Kim, Seungryong
AU - Sohn, Kwanghoon
PY - 2015/1/1
Y1 - 2015/1/1
N2 - Bag-of-words (BoW) is one of the most successful methods for object categorization. This paper proposes a statistical codeword selection algorithm where the best subset is selected from the initial codewords based on the statistical characteristics of codewords. For this purpose, we defined two types of codeword-confidences: cross- and within-category confidences. The cross- and within-category confidences eliminate indistinctive codewords across categories and inconsistent codewords within each category, respectively. An informative subset of codewords is then selected based on these two codeword-confidences. The experimental evaluation for a scene categorization dataset and a Caltech-101 dataset shows that the proposed method improves the categorization performance up to 10% in terms of error rate reduction when cooperated with BoW, sparse coding (SC), and locality-constrained liner coding (LLC). Furthermore, the codeword size is reduced by 50% leading a low computational complexity.
AB - Bag-of-words (BoW) is one of the most successful methods for object categorization. This paper proposes a statistical codeword selection algorithm where the best subset is selected from the initial codewords based on the statistical characteristics of codewords. For this purpose, we defined two types of codeword-confidences: cross- and within-category confidences. The cross- and within-category confidences eliminate indistinctive codewords across categories and inconsistent codewords within each category, respectively. An informative subset of codewords is then selected based on these two codeword-confidences. The experimental evaluation for a scene categorization dataset and a Caltech-101 dataset shows that the proposed method improves the categorization performance up to 10% in terms of error rate reduction when cooperated with BoW, sparse coding (SC), and locality-constrained liner coding (LLC). Furthermore, the codeword size is reduced by 50% leading a low computational complexity.
UR - http://www.scopus.com/inward/record.url?scp=84923673668&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84923673668&partnerID=8YFLogxK
U2 - 10.1117/12.2078771
DO - 10.1117/12.2078771
M3 - Conference contribution
AN - SCOPUS:84923673668
T3 - Proceedings of SPIE - The International Society for Optical Engineering
BT - Proceedings of SPIE-IS and T Electronic Imaging - Intelligent Robots and Computer Vision XXXII
A2 - Roning, Juha
A2 - Casasent, David
PB - SPIE
ER -