Please use this identifier to cite or link to this item: http://localhost/handle/Hannan/167886
Title: Multi-Manifold Locality Graph Embedding Based on the Maximum Margin Criterion (MLGE/MMC) for Face Recognition
Authors: Minghua Wan;Zhihui Lai
Year: 2017
Publisher: IEEE
Abstract: Solving problems with small sample sizes during training for feature extraction and the dimensionality reduction method will not produce high face recognition accuracy using the locality graph embedding (LGE) algorithm. Thus, we introduced a new algorithm named &x201C;multi-manifold locality graph embedding algorithm based on the maximum margin criterion&x201D; (MLGE/MMC) by combining the ideas of the maximum margin criterion (MMC) and multiple manifolds. First, each image is divided into multiple small images; this small image configuration constitutes a manifold and multiple images constitute multiple manifolds. Second, through maximizing the inter-class distance, while minimizing the intra-class distance to find the best class projection matrix, we build the multiple manifolds inter-class scatter matrix and the multiple manifolds intra-class scatter matrix, respectively. Finally, the objective function was constructed within the framework of the MMC to find the optimal solution by iteration and the experimental results demonstrate the effectiveness of the proposed algorithm using the ORL, Yale, and AR face image databases.
Description: 
URI: http://localhost/handle/Hannan/167886
volume: 5
More Information: 9823,
9830
Appears in Collections:2017

Files in This Item:
File SizeFormat 
7932070.pdf4.51 MBAdobe PDF
Title: Multi-Manifold Locality Graph Embedding Based on the Maximum Margin Criterion (MLGE/MMC) for Face Recognition
Authors: Minghua Wan;Zhihui Lai
Year: 2017
Publisher: IEEE
Abstract: Solving problems with small sample sizes during training for feature extraction and the dimensionality reduction method will not produce high face recognition accuracy using the locality graph embedding (LGE) algorithm. Thus, we introduced a new algorithm named &x201C;multi-manifold locality graph embedding algorithm based on the maximum margin criterion&x201D; (MLGE/MMC) by combining the ideas of the maximum margin criterion (MMC) and multiple manifolds. First, each image is divided into multiple small images; this small image configuration constitutes a manifold and multiple images constitute multiple manifolds. Second, through maximizing the inter-class distance, while minimizing the intra-class distance to find the best class projection matrix, we build the multiple manifolds inter-class scatter matrix and the multiple manifolds intra-class scatter matrix, respectively. Finally, the objective function was constructed within the framework of the MMC to find the optimal solution by iteration and the experimental results demonstrate the effectiveness of the proposed algorithm using the ORL, Yale, and AR face image databases.
Description: 
URI: http://localhost/handle/Hannan/167886
volume: 5
More Information: 9823,
9830
Appears in Collections:2017

Files in This Item:
File SizeFormat 
7932070.pdf4.51 MBAdobe PDF
Title: Multi-Manifold Locality Graph Embedding Based on the Maximum Margin Criterion (MLGE/MMC) for Face Recognition
Authors: Minghua Wan;Zhihui Lai
Year: 2017
Publisher: IEEE
Abstract: Solving problems with small sample sizes during training for feature extraction and the dimensionality reduction method will not produce high face recognition accuracy using the locality graph embedding (LGE) algorithm. Thus, we introduced a new algorithm named &x201C;multi-manifold locality graph embedding algorithm based on the maximum margin criterion&x201D; (MLGE/MMC) by combining the ideas of the maximum margin criterion (MMC) and multiple manifolds. First, each image is divided into multiple small images; this small image configuration constitutes a manifold and multiple images constitute multiple manifolds. Second, through maximizing the inter-class distance, while minimizing the intra-class distance to find the best class projection matrix, we build the multiple manifolds inter-class scatter matrix and the multiple manifolds intra-class scatter matrix, respectively. Finally, the objective function was constructed within the framework of the MMC to find the optimal solution by iteration and the experimental results demonstrate the effectiveness of the proposed algorithm using the ORL, Yale, and AR face image databases.
Description: 
URI: http://localhost/handle/Hannan/167886
volume: 5
More Information: 9823,
9830
Appears in Collections:2017

Files in This Item:
File SizeFormat 
7932070.pdf4.51 MBAdobe PDF