Learning Boltzmann distance metric for face recognition
We introduce a new method for face recognition using a versatile probabilistic model known as Restricted Boltzmann Machine (RBM). In particular, we propose to regularise the standard data likelihood learning with an information-theoretic distance metric defined on intra-personal images. This results...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Conference Paper |
| Published: |
IEEE
2012
|
| Online Access: | http://hdl.handle.net/20.500.11937/4527 |
| _version_ | 1848744541222338560 |
|---|---|
| author | Tran, Truyen Phung, D. Venkatesh, S. |
| author2 | Not known |
| author_facet | Not known Tran, Truyen Phung, D. Venkatesh, S. |
| author_sort | Tran, Truyen |
| building | Curtin Institutional Repository |
| collection | Online Access |
| description | We introduce a new method for face recognition using a versatile probabilistic model known as Restricted Boltzmann Machine (RBM). In particular, we propose to regularise the standard data likelihood learning with an information-theoretic distance metric defined on intra-personal images. This results in an effective face representation which captures the regularities in the face space and minimises the intra-personal variations. In addition, our method allows easy incorporation of multiple feature sets with controllable level of sparsity. Our experiments on a high variation dataset show that the proposed method is competitive against other metric learning rivals. We also investigated the RBM method under a variety of settings, including fusing facial parts and utilizing localised feature detectors under varying resolutions. In particular, the accuracy is boosted from 71.8% with the standard whole-face pixels to 99.2% with combination of facial parts, localised feature extractors and appropriate resolutions. |
| first_indexed | 2025-11-14T06:03:06Z |
| format | Conference Paper |
| id | curtin-20.500.11937-4527 |
| institution | Curtin University Malaysia |
| institution_category | Local University |
| last_indexed | 2025-11-14T06:03:06Z |
| publishDate | 2012 |
| publisher | IEEE |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | curtin-20.500.11937-45272017-09-13T14:44:36Z Learning Boltzmann distance metric for face recognition Tran, Truyen Phung, D. Venkatesh, S. Not known We introduce a new method for face recognition using a versatile probabilistic model known as Restricted Boltzmann Machine (RBM). In particular, we propose to regularise the standard data likelihood learning with an information-theoretic distance metric defined on intra-personal images. This results in an effective face representation which captures the regularities in the face space and minimises the intra-personal variations. In addition, our method allows easy incorporation of multiple feature sets with controllable level of sparsity. Our experiments on a high variation dataset show that the proposed method is competitive against other metric learning rivals. We also investigated the RBM method under a variety of settings, including fusing facial parts and utilizing localised feature detectors under varying resolutions. In particular, the accuracy is boosted from 71.8% with the standard whole-face pixels to 99.2% with combination of facial parts, localised feature extractors and appropriate resolutions. 2012 Conference Paper http://hdl.handle.net/20.500.11937/4527 10.1109/ICME.2012.131 IEEE fulltext |
| spellingShingle | Tran, Truyen Phung, D. Venkatesh, S. Learning Boltzmann distance metric for face recognition |
| title | Learning Boltzmann distance metric for face recognition |
| title_full | Learning Boltzmann distance metric for face recognition |
| title_fullStr | Learning Boltzmann distance metric for face recognition |
| title_full_unstemmed | Learning Boltzmann distance metric for face recognition |
| title_short | Learning Boltzmann distance metric for face recognition |
| title_sort | learning boltzmann distance metric for face recognition |
| url | http://hdl.handle.net/20.500.11937/4527 |