Vizualne tehnike analize človeškega gibanja in njihova uporaba v biometriji

doktorska disertacija

Jure Kovač (Author), Peter Peer (Mentor)

Abstract

Ljudje smo sposobni prepoznati druge ljudi, ki jih dobro poznamo, samo po tem, kako se ti premikajo. Ta sposobnost predstavlja osnovno motivacijo za uporabo načina gibanja kot sredstva za biometrično identifikacijo. Takšno biometriko lahko zajamemo na daljavo, na javnih mestih, brez posameznikovega sodelovanja, zavedanja ali celo privoljenja. Kljub temu, da današnji pristopi dajejo vzpodbudne rezultate, smo še vedno daleč od učinkovite rabe v vsakdanjih aplikacijah. V splošnem je treba upoštevati različne omejitve, da lahko metode zaobidejo številne dejavnike vpliva, kot so spremembe hitrosti hoje, pogleda, obleke, obutve in nošenja predmetov, ki imajo negativen vpliv na učinkovitost razpoznavanja. V predloženem delu predlagamo sistem za razpoznavanje ljudi na osnovi modela skeleta človeškega telesa, pri čemer se osredotočamo predvsem na modeliranje dinamike gibanja in izločanje vpliva posameznikovega videza na razpoznavanje. Predlagani metodi za klasifikacijo s fuzijo značilk in klasifikacijo na osnovi okvirjev demonstrirata primerljivost razvitega sistema z drugimi sorodnimi metodami, ki sicer temeljijo na videzu, celo v takšnih okoljih, kjer imajo slednje prednost. Dodatno naslovimo problem spremembe hitrosti hoje in s predlagamo transformacijo prostora značilk omilimo njegov negativen vpliv na učinkovitost razpoznavanja. Ta daje v kombinaciji s klasifikacijo na osnovi okvirjev celo boljše rezultate razpoznavanja kot druge - najboljše (angl. state of the art) - metode. S podrobno evalvacijo prikažemo učinkovitost in robustnost delovanja predlaganih metod tako pri nespremenjenih pogojih kot pri spremembah hitrosti hoje ter pri prisotnosti različnih ovir značilnih za širše področje računalniškega vida (npr. šum, nizka ločljivost ipd.), ki v splošnem ovirajo učinkovitost delovanja aplikacij s področja računalniškega vida.

Keywords

model skeleta;neodvisnost od hitrosti hoje;razpoznavanje ljudi;računalniški vid;biometrija;računalništvo;disertacije;

Data

Language:	Slovenian
Year of publishing:	2014
Typology:	2.08 - Doctoral Dissertation
Organization:	UL FRI - Faculty of Computer and Information Science
Publisher:	[J. Kovač]
UDC:	004.93(043.3)
COBISS:	1536014531
Views:	83
Downloads:	2
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Vision Based Techniques for Human Gait Analysis and Their use in Biometrics
Secondary abstract:	Humans are able to recognize small number of people they know well by the way they walk. This ability represents basic motivation for using human gait as the means for biometric identification. Such biometric can be captured at public places from a distance without subject’s collaboration, awareness and even consent. Although current approaches give encouraging results, we are still far from effective use in real-life applications. In general, methods set various constraints to circumvent the influence of covariate factors like changes of walking speed, view, clothing, footwear, object carrying, that have negative impact on recognition performance. In this thesis we propose a skeleton model based gait recognition system focusing on modeling gait dynamics and eliminating the influence of subject appearance on recognition. We propose feature fusion scheme for classification and frame based classification which both demonstrate how such system is comparable to other state of the art appearance based methods even in the environments where these have clear advantage. Furthermore, we address the problem of walking speed variation and propose feature space transformation technique that mitigates its negative influence on recognition performance. Together with frame based classification, the method achieves state of the art classification results and outperforms other similar methods. With the extensive evaluation we demonstrate state of the art performance and robustness of proposed methods under unchanged conditions, varying walking speeds and even under presence of different computer vision related obstacles (e.g. noise, low resolution etc.) that obstruct the performance of real-life computer vision applications in general.
Secondary keywords:	skeleton model;walking speed invariance;gait recognition;computer vision;biometrics;doctoral dissertations;theses;
File type:	application/pdf
Type (COBISS):	Doctoral dissertation
Thesis comment:	Univ. v Ljubljani, Fak. za računalništvo in informatiko
Pages:	IX, 92 str.
ID:	24252633