Skip to main content
WorldCist'17 - 5th World Conference on Information Systems and Technologies

Full Program »

Algorithms for people recognition in digital images: A systematic review and testing

People recognition in digital images has wide applications and challenges. In this article, we present a systematic review of works published in the last decade; based on which, we have identified, implemented and tested the frequently used and best-assessed algorithms. We have found Histograms of Oriented Gradients (HOG) like feature extraction algorithm; and two classification algorithms, AdaBoost and Support Vector Machine (SVM). The tests were performed on 50 images chosen randomly from Penn-Fudan public database. The accuracy in SVM-HOG combination was 0.96, it is a similar value to a related work; and the detection rate was 0.66 in SVM-HOG combination and 0.72 in Adaboost-HOG combination, they are inferior to related works. We shall discuss possible reasons.

Author(s):

Monserrate Intriago-Pazmiño    
Departamento de Informática y Ciencias de la Computación, Escuela Politécnica Nacional
Ecuador

Vanessa Vargas-Sandoval    
Escuela Politécnica Nacional
Ecuador

Jorge Moreno-Díaz    
Escuela Politécnica Nacional
Ecuador

Elizabeth Salazar-Jácome    
Departamento de Ciencias Exactas, Universidad de las Fuerzas Armadas ESPE ID: 60104598
Ecuador

Mayra Salazar-Grandes    
Departamento de Eléctrica y Electrónica, Universidad de las Fuerzas Armadas ESPE ID: 60104598
Ecuador

 

Powered by OpenConf®
Copyright ©2002-2016 Zakon Group LLC