|Author:||Pong, Kuong Hon|
|Title:||Efficient techniques for human face recognition|
|Subject:||Human face recognition (Computer science)|
Hong Kong Polytechnic University -- Dissertations
|Department:||Faculty of Engineering|
|Pages:||xvii, 130 leaves : ill. (some col.) ; 30 cm.|
|Abstract:||In this thesis, we investigate different efficient algorithms for face recognition, which has a wide range of applications. We first provide an overview of existing face recognition algorithms. Then, we devise efficient algorithms for face recognition, in particular when the input face image is of low resolution. For face recognition, image resolution is an important factor that has a great influence on the performance of a face recognition algorithm. In video surveillance, the face images concerned are usually of very low resolution. A traditional approach for low-resolution face recognition is to perform face interpolation or super-resolution, and then extract the useful features from the interpolated or super-resolved face images for recognition. To achieve a more efficient and accurate approach for low-resolution face recognition, we propose in this thesis a new method, namely "Gabor-Feature Hallucination", which predicts high-resolution Gabor features from the corresponding low-resolution Gabor features directly, by means of linear regression and Generalized Canonical Correlation Analysis. Then, the low-resolution features in the projected Generalized Canonical Correlation space and the predicted high-resolution Gabor features are adopted for face classification. This algorithm can avoid having to perform interpolation/super-resolution and then high-resolution Gabor feature extraction. Our algorithm estimates the high-resolution features, which are combined with the original low-resolution features to form an efficient feature using Canonical Correlation Analysis. Experimental results show that the proposed approach has a superior recognition rate and efficiency compared to the traditional methods.|
|Rights:||All rights reserved|
Files in This Item:
|b25545978.pdf||For All Users (off-campus access for PolyU Staff & Students only)||2.42 MB||Adobe PDF||View/Open|
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item: