Navigation

Recent Publications

More...

Home

Hidden Markov Models for Face Recognition

u3003097 - Posted on 05 April 2016

Project Description: 

Using eye-tracking data, we aim to build a Hidden Markov Model for each participant to describe spatial and temporal aspects of the participant’s face recognition strategy. Moreover,we are currently exploring different HMM architectures (e.g. HMM with image information).

Researcher name: 
Janet Hsiao
Researcher position: 
Associate Professor
Researcher department: 
Department of Psychology
Researcher email: 
jhsiao@hku.hk
Research Project Details
Project Duration: 
01/2015 - 12/2017
Project Significance: 
The Hidden Markov Model which is learned using fixation data (and possibly image information) helps researchers to better model spatial and temporal aspects of eye-tracking data of various cognitive tasks.
Results Achieved: 
We will update the 'Results Achieved' section once we have published our research.
Remarks: 
The HPC substantially helps us to speed up and parallelize our analyses. To accurately identify important regions of interests on face stimuli, we need to perform many fitting iterations and select the global optimum. The latter is computationally very expensive.