Efeitos farmacológicos da hidroxicloroquina em pacientes com COVID-19: Revisão sistemática da literatura
Year: 2019
Type: article
Abstract: Texture pattern observed on finger knuckle joint is called as Finger Knuckle Print (FKP). FKP can be extracted from inner side or back side of finger surface. FKP is extremely unique and makes the knuckle surface an emerging biometric modality. The FKP may be useful in user identification and has attracted attention of very few researchers. Our proposed work is based on FKP of back side of finger surface. The objective of this work is to investigate and develop a systematic approach for identifying a person using his/her finger knuckle print. Three different methodologies are employed for FKP identification process. Statistical approach (Principal Component Analysis) and feature based approaches (Gabor based coding scheme and Radon like features) are used to accomplish this objective. Nearest Mean Classification, Angular distance matching and Dynamic Time Warping methods are used for finding similarity between enrolled FKP and test FKP images. Performance comparison of proposed three methodologies is done by computing different performance measures such as FAR, FRR, EER, Di and CRR. Proposed methods are producing EER 0.2 using statistical approach and 0.03 using feature based approach.
Authors: Francisco Emanoel Alves de Araújo, João Matheus Caé da Rocha, Lucas Emmanuel Rocha de Moura Marques, Kellyson Lopes da Silva Macedo, Paloma Katlheen Moura Melo, Gabriella Mendes Duarte, Francisco Irochima Pinheiro, Fausto Pierdoná Guzen
Language: English
Subfield: Signal Processing
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W3178564677