PCA-AdaBoost-LDA Face Recognition Algorithm

This book presents a new method to recognize faces with high accuracy for the above aspects. A method with 68 points landmark-based face estimation and image normalization with AdaBoost-LDA for poses and illumination invariant face recognition is proposed. A single training image per person is derived from number of training image samples using average intensity values to reduce memory and execution time. AdaBoost-LDA is used for extraction of feature and classic nearest centre classifier is used for feature classification. Proposed method has successfully handled the illumination conditions, pose variations, and occlusion in low resolution images. Experimental results illustrate the promising performance of presented approach over the current published approaches on LFW, AR and CMU Multi-PIE databases.

Autorentext

Mahmood Ul Haq received a B.S. degree in Electrical and Electronics Engineering from COMSATS University Islamabad (CUI), Abbottabad Campus, Pakistan, in 2016. He received his MS Electrical Engineering in COMSATS University Islamabad (CUI), Pakistan. Mahmood's research expertise encompasses topics, such as Pattern Recognition and Machine Learning.