Please use this identifier to cite or link to this item:
http://hdl.handle.net/10603/13445
Title: | An improved multiple weighted facial attribute set based face recognition system using statistical and machine learning techniques |
Researcher: | Sakthivel S |
Guide(s): | Rajaram, M. |
Keywords: | Statistical techniques, machine learning techniques, Multiple Weighted Facial Attribute Sets |
Upload Date: | 28-Nov-2013 |
University: | Anna University |
Completed Date: | |
Abstract: | In recent years, research on face recognition has attracted more and more attention from both academia and industry. Face Recognition has become an important concept in many applications which is wide open that covers amongst other security features. Face Recognition addresses several problems in pose, expression, illumination, age, lightning and alignment. While distinguishing a particular face from all other faces in any recognition system, the above difficulty problem was raised. After a comprehensive study on the existing face recognition techniques and doing an evaluation on feature extraction and dimensionality reduction techniques, a most suitable technique is used during designing the proposed face recognition system. This work describes a new approach to a face recognition system for image querying in image database applications. If an input face image and a probable database are given, a set of possible candidates will be found. It is subjected to the constraint that the faces are matched from the input image A new technique, Multiple Weighted Facial Attribute Sets integrated with PCA algorithm is proposed in this thesis that describes add-on features to face recognition to improve better performance. In this research, it has been shown that the weights given for the individual attributes influenced the overall performance of the recognition system. The performance of proposed face recognition model was tested with the standard Database. The performance of the system is improved further using wavelet based image decomposition technique . A significant 8.54% performance improvement was observed during various tests. The proposed technique will have a user-defined input component which will give some priority to a set of features of the image during the matching process. This makes the proposed algorithm unique and has a positive impact on the performance. newline newline newline |
Pagination: | xiii, 133 |
URI: | http://hdl.handle.net/10603/13445 |
Appears in Departments: | Faculty of Information and Communication Engineering |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
01_title.pdf | Attached File | 76.45 kB | Adobe PDF | View/Open |
02_certificates.pdf | 1 MB | Adobe PDF | View/Open | |
03_abstract.pdf | 72.85 kB | Adobe PDF | View/Open | |
04_acknowledgement.pdf | 83.53 kB | Adobe PDF | View/Open | |
05_contents.pdf | 101.75 kB | Adobe PDF | View/Open | |
06_chapter 1.pdf | 195.74 kB | Adobe PDF | View/Open | |
07_chapter 2.pdf | 280.5 kB | Adobe PDF | View/Open | |
08_chapter 3.pdf | 214.05 kB | Adobe PDF | View/Open | |
09_chapter 4.pdf | 340.59 kB | Adobe PDF | View/Open | |
10_chapter 5.pdf | 215.06 kB | Adobe PDF | View/Open | |
11_chapter 6.pdf | 252.51 kB | Adobe PDF | View/Open | |
12_chapter 7.pdf | 83.78 kB | Adobe PDF | View/Open | |
13_references.pdf | 114.87 kB | Adobe PDF | View/Open | |
14_publications.pdf | 85.92 kB | Adobe PDF | View/Open | |
15_vitae.pdf | 75.96 kB | Adobe PDF | View/Open |
Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).
Altmetric Badge: