Sign Language Research Paper

Abstract — Sign language is the language used by deaf and dumb to communicate among themselves and with normal people. A functioning sign language recognition (SLR) system generate speech or text provide an opportunity for a mute person to communicate with non-signing people without the need for an interpreter. The SLR system consists of acquisition of image, Segmentation, Feature extraction and classification or recognition. In this paper, we provide a survey on segmentation and feature extraction techniques for Indian Sign Language Recognition system.

Keywords— Sign languages, ISL Recognition System, Segmentation, Feature Extraction, object-based video abstraction(VOPs), Skin segmentation, orientation histogram, B-Spline curve approximation

Section III focuses on a typical sign language recognition system. Section IV describes the segmentation of vision based hand gesture recognition system with various techniques available in literature. Section V discuss about the challenges in developing the ISL recognition system. Section VI contains discussion and conclusion.

II. INDIAN SIGN LANGUAGE

Sign languages are natural languages used by deaf and dump which uses various means of expression for communication in everyday life. It relates letters, words, and sentences of a spoken language to gestures facilitating hearing impaired people to communicate among themselves. Sign languages are used not only by the deaf people but also by the hearing parents of the deaf children, hearing children of deaf adults and hearing deaf educators.

The native language commonly practiced by the deaf community of India is the Indian Sign Language. It allows a deaf to convey thoughts and ideas using hands, arms, and face. Unlike spoken languages, ISL uses gestures instead of sounds in order to express a

CHALLENGES IN SIGN LANGUAGE RECOGNITION SYSTEM

Sign language function and spoken language functions are totally different. SL is based on spatial properties and iconicity properties. The primary challenge faced by any sign language recognition system is the ability to track the signer in the video of the signer with a variety of background clutter and different lighting conditions. Britta and Karl-Friedrich [27] reported some difficulties in terms of sign language:

i)Occlusions problem while performing sign ii) Signer position may vary in front of camera while performing sign so that the camera can't acquire some extrems of the sign. iii) A 2D camera give loss of depth information iv) As each sign varies in time and space, so there may be a change in position and speed with same person or person to person

v)Co-articulation problem (link between preceding and subsequent sign)

VI. CONCLUSIONS

The major objective of this paper was to study the various methods of segmentation and feature extraction used for ISL recognition system. Most of the methods developed so far use simple backgrounds in controlled set-up, special hardware like data gloves, restricted sets of actions, and restricted number of