Feature Extraction Methods in Sign Language Recognition System: A Literature Review

摘要

Sign language is a way to communicate between deaf-mute people and normal people by performing hand gesture. Visual-based gesture recognition can help to overcome this communication limitation. Recently, several techniques and methods have been proposed in this area of research and showed some improvements. Despite all of the proposed methods, most of hand gesture recognition approaches that have been applied still lack of compatibility and have lots of limitations. For instance, hand segmentation meets the complication of distinguishing the hand and the face region by using skin detection. Motivated by those facts, this paper presents a review and explains progress of feature extraction in sign language recognition mostly in the last ten years. In this contribution, we focus on studying feature extraction methods. The literature used in this study is based on the previous published international papers which discussed sign language recognition. The main objectives from this review are to get the most effective and most compatible feature extraction method to be applied to sign language recognition system and to further research progress in the future. After reviewing various recognition techniques, we can conclude that the current works have successfully improve hand gesture recognition by inventing a technology which helps for tracking hands region precisely by using an active sensor. However, there is still room for improvements based on a markerless passive sensor, such as vision-based approaches.