Abstract—Lip tracking is very crucial for visual lip reading recognition system. This paper presents a novel active contour guided geometrical feature extraction approach for lip reading. Three active contour methods are studied, namely snake, region scalable fitting energy method and loacalised active contour model. These methods are adopted for salient geometrical feature calculation. A joint feature model, obtained by combinatining inner area, height and width has been proposed. Results of experimentations on digit utterances are given to show the improvement achieved by visual speech recognition systems.
Index Terms—Localised active contour model, lip tracking, principal component analysis, region scalable fitting energy method.
Sunil S. Morade is with the Department of E and TC Engineering, K K Wagh Institute of Engineering Education & Research, Nashik, India (e-mail: ssm.eltx@gmail.com).
B. Suparva Patnaik was with SVNIT, Surat, India. She is now with the Department of E and TC Engineering, Xavier Institute of Engineering, Mumbai, India (e-mail: suprava_patnaik@yahoo.com).
Cite:Sunil S. Morade and B. Suprava Patnaik, "Automatic Lip Tracking and Extraction of Lip Geometric Features for Lip Reading," International Journal of Machine Learning and Computing vol. 3, no. 2, pp. 168-171, 2013.