Please use this identifier to cite or link to this item:
http://hdl.handle.net/10603/355051
Title: | Face Recognition From Low Resolution Images |
Researcher: | RENJITH THOMAS |
Guide(s): | Rangachar, M J S |
Keywords: | Engineering Engineering and Technology Engineering Electrical and Electronic |
University: | Hindustan University |
Completed Date: | 2019 |
Abstract: | Face recognition has been a major research interest due to its applicability in newlinevarious fields. For better face recognition, inter-personal and intra-personal newlinevariations are to be considered. The purpose of this research is to overcome the newlinepractical difficulties met during the recognition of face from low resolution newlineimages. Hence, this research addresses these issues and propose five techniques newlinefor effective face recognition. The first contribution is the recognition newlineperformed by the extraction of features using Gabor filters. A large set of newlinefeatures is obtained by the use of Gabor filter and multi-lobe differential filters. newlineThese features are then fed to dimensionality reduction algorithm. Here sparse newlineprincipal component analysis (SPCA) is used for dimensionality reduction of newlinethe feature set. A Fuzzy logic classifier is used for classification. The second newlinecontribution is a design of new feature extraction technique known as Gabor newlinewavelet texture model (GWTM). Features extracted from the model are then newlineundergone fusion technique. An optimum fusion parameter is obtained by the newlineuse of an optimization algorithm called Bat algorithm. Classifier used in the newlinework is Spherical support vector machine (SSVM), which can classify features newlineefficiently. The third contribution is the use of GWTM for feature extraction newlineand a new optimization algorithm called fractional Bat, which incorporates newlinefractional theory. The fourth contribution is the designing a face recognition newlinesystem with the combination of GWTM operator and the crow search algorithm newlinefor increasing the performance of classification. The fifth contribution is a new newlineclassifier called Multi-kernel spherical support vector machine with GWTM newlineoperator and the whale optimization algorithm, which can handle non-linear newlinedata more efficiently. The evaluation metrics used are false acceptance rate newline(FAR), False rejection rate (FRR) and accuracy. |
Pagination: | |
URI: | http://hdl.handle.net/10603/355051 |
Appears in Departments: | Department of Electronics and Communication Engineering |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
10_chapter 2.pdf | Attached File | 703.89 kB | Adobe PDF | View/Open |
11_chapter 3.pdf | 2.87 MB | Adobe PDF | View/Open | |
12_chapter 4.pdf | 171.03 kB | Adobe PDF | View/Open | |
13_chapter 5.pdf | 377.94 kB | Adobe PDF | View/Open | |
14_chapter 6.pdf | 373.24 kB | Adobe PDF | View/Open | |
15_bibliography.pdf | 521.72 kB | Adobe PDF | View/Open | |
1_title.pdf | 281.08 kB | Adobe PDF | View/Open | |
2_certificate.pdf | 1.37 MB | Adobe PDF | View/Open | |
3_declaration.pdf | 282.15 kB | Adobe PDF | View/Open | |
4_acknowledgment.pdf | 364.93 kB | Adobe PDF | View/Open | |
5_table of contents.pdf | 534.13 kB | Adobe PDF | View/Open | |
6_abstract.pdf | 345.7 kB | Adobe PDF | View/Open | |
7_list of tables.pdf | 529.69 kB | Adobe PDF | View/Open | |
80_recommendation.pdf | 11.8 MB | Adobe PDF | View/Open | |
8_symbols and abbreviations.pdf | 621.13 kB | Adobe PDF | View/Open | |
9_chapter 1.pdf | 1.56 MB | 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: