Please use this identifier to cite or link to this item: http://hdl.handle.net/2445/48865
Title: Multi-clasificación discriminativa de partes corporales basada en códigos correctores de errores
Author: Pérez Yarza, José Tomás
Director: Escalera Guerrero, Sergio
Bautista Martín, Miguel Ángel
Keywords: Visió per ordinador
Reconeixement de formes (Informàtica)
Programari
Tesis
Computer vision
Pattern recognition systems
Computer software
Theses
Issue Date: 20-Sep-2013
Abstract: This Project aims at the application of different techniques from the field of Artificial Vision for the detection and segmentation of human limbs on a newly created database. The database contains a large number of images where multiple subjects appear performing various poses. The objective is to detect the limbs of such subjects, including the arms, legs, body or head, among others, to subsequently obtain a multi-limb segmentation map. In order to perform this detection we trained different classifiers cascades on Haar and HOG features on targeted regions where limbs appeared. Once trained, several experiments have been released over the database for detecting the limbs mentioned. Some methods have been used to verify the detections. Finally, segmentation techniques have been applied for two purposes: On one hand, segment the subject from the background of the image, and on the other hand, each limb of the subject. In this case we have chosen segmentation using Graph-cuts formulation.
Note: Treballs Finals de Grau d'Enginyeria Informàtica, Facultat de Matemàtiques, Universitat de Barcelona, Any: 2013, Director: Sergio Escalera Guerrero i Miguel Ángel Bautista Martín
URI: http://hdl.handle.net/2445/48865
Appears in Collections:Treballs Finals de Grau (TFG) - Enginyeria Informàtica
Programari - Treballs de l'alumnat

Files in This Item:
File Description SizeFormat 
memoria.pdfMemòria2.37 MBAdobe PDFView/Open
src.zipCodi Font757.6 kBzipView/Open


This item is licensed under a Creative Commons License Creative Commons