4.3 Article

Analysis of facial expressions in Indian histrionics using a Petri-net verified process and a state model

Journal

Publisher

SPRINGER INDIA
DOI: 10.1007/s12046-022-02030-x

Keywords

Facial expression analysis; Petri-nets; CPN tool; e-tutor for facial expressions; Indian histrionics

Ask authors/readers for more resources

Facial expressions are crucial in theatrical performances, and this article introduces a facial expression analysis system that is based on the nine specific expressions described in ancient Indian literature. The system uses a combination of facial features and states to identify expressions and provides feedback to performers through an e-tutor system, helping them improve their expression skills.
Facial expressions are a very important aspect of a theatrical performance and analysis of expressions used in histrionics is an unexplored application. The systems for facial expression analysis typically analyse SIX basic expressions viz. happiness, sadness, anger, surprise, fear, and disgust. The developed system proposes to analyse NINE expressions that are specific to Indian histrionics and are described in ancient literature as a combination of relevant facial feature states. In this respect, the feature state approach using six Euclidean distance measures is designed to ensure the individual contributions of facial components during classification. To make the system invariant to the original facial features and to invalidate the effect of the same, a subject-dependent approach is used in this study. This approach compares neutral and expressive images of the same subject based on distance measures to determine the feature state. The process of expression recognition using facial component states is modelled and simulated using a Coloured Petri Net (CPN) tool. A novel state model is designed for the identification of expressions using a minimum number of states. A database that includes facial images exhibiting nine performance-specific expressions is created. Upon testing, the user either receives information about correct expressions or about incorrectly positioned facial states which then can be used for improvisation. This way, the proposed system acts as an e-tutor and helps the performers for the betterment of their expression skills irrespective of the classification result. The proposed system is tested on 1440 samples from the dataset. For 1190 samples, all the feature states are found to be correct, and the image is labelled with an expression. For the rest 250 samples, the e-tutor system provided information about the incorrect feature states. The person-dependent method allows the system to be used for global users irrespective of their age, gender, and ethnicity.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.3
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available