Robust Single-Sample Face Recognition by Sparsity-Driven Sub-Dictionary Learning Using Deep Features
Face recognition using a single reference image per subject is challenging, above all when referring to a large gallery of subjects. Furthermore, the problem hardness seriously increases when the images are acquired in unconstrained conditions. In this paper we address the challenging Single Sample...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2019-01-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | http://www.mdpi.com/1424-8220/19/1/146 |
| _version_ | 1798024668465070080 |
|---|---|
| author | Vittorio Cuculo Alessandro D’Amelio Giuliano Grossi Raffaella Lanzarotti Jianyi Lin |
| author_facet | Vittorio Cuculo Alessandro D’Amelio Giuliano Grossi Raffaella Lanzarotti Jianyi Lin |
| author_sort | Vittorio Cuculo |
| collection | DOAJ |
| description | Face recognition using a single reference image per subject is challenging, above all when referring to a large gallery of subjects. Furthermore, the problem hardness seriously increases when the images are acquired in unconstrained conditions. In this paper we address the challenging Single Sample Per Person (SSPP) problem considering large datasets of images acquired in the wild, thus possibly featuring illumination, pose, face expression, partial occlusions, and low-resolution hurdles. The proposed technique alternates a sparse dictionary learning technique based on the method of optimal direction and the iterative ℓ 0 -norm minimization algorithm called k-LiMapS. It works on robust deep-learned features, provided that the image variability is extended by standard augmentation techniques. Experiments show the effectiveness of our method against the hardness introduced above: first, we report extensive experiments on the unconstrained LFW dataset when referring to large galleries up to 1680 subjects; second, we present experiments on very low-resolution test images up to 8 × 8 pixels; third, tests on the AR dataset are analyzed against specific disguises such as partial occlusions, facial expressions, and illumination problems. In all the three scenarios our method outperforms the state-of-the-art approaches adopting similar configurations. |
| first_indexed | 2024-04-11T18:06:23Z |
| format | Article |
| id | doaj.art-67223279e17847e582d1904b53158235 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-04-11T18:06:23Z |
| publishDate | 2019-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-67223279e17847e582d1904b531582352022-12-22T04:10:19ZengMDPI AGSensors1424-82202019-01-0119114610.3390/s19010146s19010146Robust Single-Sample Face Recognition by Sparsity-Driven Sub-Dictionary Learning Using Deep FeaturesVittorio Cuculo0Alessandro D’Amelio1Giuliano Grossi2Raffaella Lanzarotti3Jianyi Lin4Dipartimento di Informatica, Università degli Studi di Milano, via Celoria 18, 20133 Milano, ItalyDipartimento di Informatica, Università degli Studi di Milano, via Celoria 18, 20133 Milano, ItalyDipartimento di Informatica, Università degli Studi di Milano, via Celoria 18, 20133 Milano, ItalyDipartimento di Informatica, Università degli Studi di Milano, via Celoria 18, 20133 Milano, ItalyDepartment of Mathematics, Khalifa University of Science and Technology, Al Saada Street, PO Box 127788, Abu Dhabi, UAEFace recognition using a single reference image per subject is challenging, above all when referring to a large gallery of subjects. Furthermore, the problem hardness seriously increases when the images are acquired in unconstrained conditions. In this paper we address the challenging Single Sample Per Person (SSPP) problem considering large datasets of images acquired in the wild, thus possibly featuring illumination, pose, face expression, partial occlusions, and low-resolution hurdles. The proposed technique alternates a sparse dictionary learning technique based on the method of optimal direction and the iterative ℓ 0 -norm minimization algorithm called k-LiMapS. It works on robust deep-learned features, provided that the image variability is extended by standard augmentation techniques. Experiments show the effectiveness of our method against the hardness introduced above: first, we report extensive experiments on the unconstrained LFW dataset when referring to large galleries up to 1680 subjects; second, we present experiments on very low-resolution test images up to 8 × 8 pixels; third, tests on the AR dataset are analyzed against specific disguises such as partial occlusions, facial expressions, and illumination problems. In all the three scenarios our method outperforms the state-of-the-art approaches adopting similar configurations.http://www.mdpi.com/1424-8220/19/1/146face recognitionsingle sample per persondictionary learningoptimal directions (MOD)Deep Convolutional Neural Network (DCNN) featuressparse recovery |
| spellingShingle | Vittorio Cuculo Alessandro D’Amelio Giuliano Grossi Raffaella Lanzarotti Jianyi Lin Robust Single-Sample Face Recognition by Sparsity-Driven Sub-Dictionary Learning Using Deep Features Sensors face recognition single sample per person dictionary learning optimal directions (MOD) Deep Convolutional Neural Network (DCNN) features sparse recovery |
| title | Robust Single-Sample Face Recognition by Sparsity-Driven Sub-Dictionary Learning Using Deep Features |
| title_full | Robust Single-Sample Face Recognition by Sparsity-Driven Sub-Dictionary Learning Using Deep Features |
| title_fullStr | Robust Single-Sample Face Recognition by Sparsity-Driven Sub-Dictionary Learning Using Deep Features |
| title_full_unstemmed | Robust Single-Sample Face Recognition by Sparsity-Driven Sub-Dictionary Learning Using Deep Features |
| title_short | Robust Single-Sample Face Recognition by Sparsity-Driven Sub-Dictionary Learning Using Deep Features |
| title_sort | robust single sample face recognition by sparsity driven sub dictionary learning using deep features |
| topic | face recognition single sample per person dictionary learning optimal directions (MOD) Deep Convolutional Neural Network (DCNN) features sparse recovery |
| url | http://www.mdpi.com/1424-8220/19/1/146 |
| work_keys_str_mv | AT vittoriocuculo robustsinglesamplefacerecognitionbysparsitydrivensubdictionarylearningusingdeepfeatures AT alessandrodamelio robustsinglesamplefacerecognitionbysparsitydrivensubdictionarylearningusingdeepfeatures AT giulianogrossi robustsinglesamplefacerecognitionbysparsitydrivensubdictionarylearningusingdeepfeatures AT raffaellalanzarotti robustsinglesamplefacerecognitionbysparsitydrivensubdictionarylearningusingdeepfeatures AT jianyilin robustsinglesamplefacerecognitionbysparsitydrivensubdictionarylearningusingdeepfeatures |