Pregled bibliografske jedinice broj: 899183
Face Detection Based on Invariant Moments Classified by Neural network
Face Detection Based on Invariant Moments Classified by Neural network // INTERNATIONAL JOURNAL OF CIRCUITS, SYSTEMS AND SIGNAL PROCESSING, 8 (2014), 203-210 (međunarodna recenzija, članak, znanstveni)
CROSBI ID: 899183 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Face Detection Based on Invariant Moments
Classified by Neural network
Autori
Capor Hrosik, Romana ; Tuba, Milan ; Vukovic, Mirjana ; Pikula, Miljenko
Izvornik
INTERNATIONAL JOURNAL OF CIRCUITS, SYSTEMS AND SIGNAL PROCESSING (1998-4464) 8
(2014);
203-210
Vrsta, podvrsta i kategorija rada
Radovi u časopisima, članak, znanstveni
Ključne riječi
Image processing, face detection, skin color detection, Hu’s moments, neural networks
(Image processing, face detection, skin color detection, Hu’s moments, neural networks.)
Sažetak
Face detection is very important topic in image processing. In this paper we propose a method for face detection based on skin detection for initial potential regions detection and Hu invariant moments as region descriptors. Seven Hu moments present complex relations among different skin regions and neural network is used for face or non-face classification. The utilized network is a multilayer perceptron (MLP) with one hidden layer. The back propagation learning is used for its training. Experimental results demonstrated successful face detection.
Izvorni jezik
Engleski
Znanstvena područja
Matematika, Računarstvo
