000			02780nam a22002417a 4500
003			FT8886
005			20251217151545.0
008			251217b ||||| |||| 00| 0 eng d
041			_aengtag
050			_aQA76.9 A43 C37 2025
082			_a.
100	1		_aCarison, Ron Hale I.; Upaga, Chloe Gwyneth S.
245			_aAn enhancement of eigenface algorithm applied for identifying spoofing attacks in facial recognition
264		1	_3. _a. _b. _cc2025
300			_bUndergraduate Thesis: (Bachelor of Science in Computer Science) - Pamantasan ng Lungsod ng Maynila, 2025
336			_2text _atext _btext
337			_2unmediated _aunmediated _bunmediated
338			_2volume _avolume _bvolume
505			_aABSTRACT: The Eigenface Algorithm uses principal component analysis (PCA) to extract and represent facial features as eigenvectors for comparison with stored datasets. This is used for facial recognition that identifies individuals by extracting facial features and converting them into eigenvectors for comparison with stored datasets. However, it struggles with face occlusions, low-resolution images, and varying distances, affecting accuracy and increasing vulnerability to spoofing attacks. This research enhances the Eigenface Algorithm by integrating Super Resolution for improved facial feature extraction, LBPH for better occlusion and spoof detection, and a Distance-Based Scaling factor to optimize recognition within a 30 cm to 60 cm range. In this research, the researchers implemented OpenCV2 and stored the trained dataset in a YAML file. The dataset was generated by capturing multiple images across different environments and distances. The facial images were then preprocessed by resizing, converting to grayscale, incorporating Super-Resolution to enhance face image quality and applying LBPH for vector representation. Additionally, Distance Scaling was integrated to optimize facial recognition across varying distances. Results demonstrated improvements, with confidence levels reaching 90.79% and a substantial reduction in the error rate of 29.45%. The accuracy rate increased to 69.44%, while recognition time was remarkably decreased to just 0.0093 seconds. Additionally, the researchers recommend adapting the enhanced eigenface algorithm for other device types, 3D application and using larger dataset. These enhancements strengthen the algorithm’s resilience against spoofing attacks, augment its reliability in extracting intricate facial details amid noise, improve accuracy in recognizing occlusions and spoofed images, and enhance its usability in recognition attempts at unsuitable distances, thus enhancing usability in practical applications of facial recognition
526			_aF
655			_aacademic writing
942			_2ddc _cMS
999			_c37364 _d37364