Deep Learning Based Face Recognition with Sparse Representation Classification
| Autor: | Eric-Juwei Cheng, Deepak Puthal, Po-Hao Chin, Nabin Sharma, Om Kumar Prasad, Chin-Teng Lin, Michael Blumenstein, Mukesh Prasad |
|---|---|
| Přispěvatelé: | Liu, D, Xie, S, Li, Y, Zhao, D, ElAlfy, ESM |
| Rok vydání: | 2017 |
| Předmět: |
Computer science
business.industry Deep learning Feature extraction 020206 networking & telecommunications Pattern recognition 02 engineering and technology Sparse approximation Facial recognition system Convolutional neural network Face (geometry) Pattern recognition (psychology) 0202 electrical engineering electronic engineering information engineering Feature (machine learning) Artificial Intelligence & Image Processing 020201 artificial intelligence & image processing Computer vision Artificial intelligence business |
| Zdroj: | Neural Information Processing ISBN: 9783319700892 ICONIP (3) |
| DOI: | 10.1007/978-3-319-70090-8_67 |
| Popis: | Feature extraction is an essential step in solving real-world pattern recognition and classification problems. The accuracy of face recognition highly depends on the extracted features to represent a face. The traditional algorithms uses geometric techniques, comprising feature values including distance and angle between geometric points (eyes corners, mouth extremities, and nostrils). These features are sensitive to the elements such as illumination, variation of poses, various expressions, to mention a few. Recently, deep learning techniques have been very effective for feature extraction, and deep features have considerable tolerance for various conditions and unconstrained environment. This paper proposes a two layer deep convolutional neural network (CNN) for face feature extraction and applied sparse representation for face identification. The sparsity and selectivity of deep features can strengthen sparseness for the solution of sparse representation, which generally improves the recognition rate. The proposed method outperforms other feature extraction and classification methods in terms of recognition accuracy. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|