Estimation of Image Features Representing Facial Emotions for Emotion Synthesis.

Date of Submission

December 2015

Date of Award

Winter 12-12-2016

Institute Name (Publisher)

Indian Statistical Institute

Document Type

Master's Dissertation

Degree Name

Master of Technology

Subject Name

Computer Science


Electronics and Communication Sciences Unit (ECSU-Kolkata)


Mukherjee, Dipti Prasad (ECSU-Kolkata; ISI)

Abstract (Summary of the Work)

We develop a method to estimate emotion-specific features on human face. Application of such a method include characterizing an emotion class and synthesis of emotions. The emotionspecific features can also be used to study the statistical differences between two clusters, one facial expression images with no expressions and two facial expression images with some or maximum emotional content. Once the feature vectors are extracted from the input data, we classify the data and use the normal to the classifier to trace the changes that a facial expression image may undergo in different stages of an emotion. We use Support Vector Machines learning algorithm to construct an optimal classifier. In the result section we show that we are able to reduce the number of features by 66.36% as compared to the total number of pixels. We show that using these features and state-of-the-art methods to synthesize images, we improved SNR of the synthesized image by 13.20% and also improved the cluster measures between a cluster of no-expression images and a cluster of with-expression images.


ProQuest Collection ID:

Control Number


Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.


This document is currently not available here.