V2I12P39

Image-Based Methods for Hand Gesture Detection and Sign Language Recognition

Ajaykumar R^1*

Abstract

Sign Language serves as the primary medium of communication for individuals with hearing and speech impairments, enabling them to convey emotions and thoughts effectively. However, the lack of familiarity with sign language among the general population often necessitates the presence of interpreters to facilitate communication. Consequently, the development of an automated sign language recognition system has become a crucial research area. This paper presents a user-independent framework for the automatic recognition of two-handed dynamic sign language gestures, aiming to bridge the communication gap between signers and non-signers. The proposed methodology comprises their major phases: preprocessing, feature extraction, and recognition. In the preprocessing phase, the system extracts the most distinctive keyframes from sign videos and applies skin color detection techniques to isolate hand regions while eliminating facial areas. This step enhances system efficiency by focusing only on the most informative frames. In the feature extraction phase, feature is derived using fuzzy triangular membership functions, which capture the spatial relationship between pixels, thereby providing a robust representation of each sign. The extracted feature vectors are then organized and compared against a reference database for classification and interpretation. Experimental evaluations were conducted on a custom dataset comprising sign videos and sentences performed by multiplying signers. The dataset primarily includes two-handed dynamic signs. Comparative analysis with stored database entries demonstrates that the proposed framework achieves high recognition accuracy, validating its effectiveness and robustness in recognizing complex sign gestures. The obtained results are promising and indicate significant potential for real-time sign language interpretation applications.

Keywords:

Sign Language Recognition; Two-Handed Dynamic Gestures; Keyframe Extraction; Fuzzy Feature Extraction; Human–Computer Interaction