Bayesian face recognition using a Markov chain Monte Carlo method

Matsui, Atsushi; Clippingdale, Simon; Uzawa, Fumiki; Matsumoto, Takashi

doi:10.1109/icpr.2004.1334678

Cited by 9 publications

(3 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Unless there are reasons to do otherwise, we set the initial model probabilities uniformly: P (H j |y 0 ) = 1/N persons . Table 1 shows recognition results for the new online learning algorithm (Bayesian SMC), compared with a batch learning algorithm (Bayesian MCMC), with which we estimated the predictive distribution of parameters using a Markov Chain Monte Carlo method [7]. Template images and test sequences showed 7 Japanese actors and 3 Japanese actresses in frontal pose against a blue background.…”

Section: Sequential Importance Samplingmentioning

confidence: 99%

“…In this work, we introduce an online learning algorithm for Bayesian posterior probabilities describing faces in video input sequences, which uses a Sequential Monte Carlo (SMC) method [4] [5] to perform integrations over a sequence of combined spaces of face model parameters and system hyperparameters. We show that this SMC approach successfully adapts the parameters associated with deformations of each face model, and significantly reduces recognition errors on a video test set showing individuals talking, relative to a baseline batch MCMC (Markov Chain Monte Carlo) algorithm [6] [7]. However, it does so at increased computational cost.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Sequential Monte Carlo Method for Bayesian Face Recognition

Matsui

Clippingdale

Matsumoto

2006

Lecture Notes in Computer Science

View full text Add to dashboard Cite

This paper proposes a Sequential Monte Carlo (SMC) learning algorithm for Bayesian probability distributions that describe model parameters in a video face recognition system based on deformable template matching. The new algorithm achieves significantly improved robustness of recognition against facial expressions and speech movements by comparison with a baseline batch MCMC (Markov Chain Monte Carlo) algorithm, at no additional computational cost. Experimental results demonstrate the effectiveness and computational efficiency of the new algorithm.

show abstract

Section: Sequential Importance Samplingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A Sequential Monte Carlo Method for Bayesian Face Recognition

Matsui

Clippingdale

Matsumoto

2006

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…Additionally, the system overall was too computationally intensive to attain reasonable (near real time) speed with standard PC hardware. Some extensions of the original system to statistical matching methods (Matsui et al, 2004(Matsui et al, , 2006 and 3D registration (Clippingdale et al, 2009) offered improved performance in various respects, but at the cost of even more intensive computation.…”

Section: Introductionmentioning

confidence: 99%

Video Face Tracking and Recognition with Skin Region Extraction and Deformable Template Matching

Clippingdale¹,

Fujii²

2012

International Journal of Multimedia Data Engineering and Management

View full text Add to dashboard Cite

The authors describe a face tracking and recognition system for video indexing that handles variable face poses (left-right and up-down) and deformations due to speech and facial expressions. The system is based on deformable template matching, and employs person-specific templates at near-frontal poses for recognition, and novel person-independent templates at multiple poses on the view-sphere for tracking. Relative to an earlier version that used multiple person-specific templates at multiple (left-right) poses, the new system speeds up processing by (i) restricting attention to skin-color regions; (ii) performing recognition using the person-specific templates at near-frontal poses only; and (iii) tracking at non-frontal poses using the novel person-independent templates. Registration is also simplified since multiple views of each target individual are no longer required, but at the cost of a loss of recognition functionality at poses far from frontal (the system instead “remembers” the identity of each individual from near-frontal matches and tracks between them).

show abstract

Identity Management in Face Recognition Systems

Tistarelli

Grosso

2008

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Bayesian face recognition using a Markov chain Monte Carlo method

Cited by 9 publications

References 5 publications

A Sequential Monte Carlo Method for Bayesian Face Recognition

A Sequential Monte Carlo Method for Bayesian Face Recognition

Video Face Tracking and Recognition with Skin Region Extraction and Deformable Template Matching

Identity Management in Face Recognition Systems

Contact Info

Product

Resources

About