WCSE

Gesture Recognition Using Graph Method

Abstract— Recognition of gestures is a challenging task that often plays an important role in enabling smooth interaction among people and between people and new age devices. Inclusion of gesture-based device interactions allows intuitive usage of technology and ease of communication. The paper proposes a new and under-explored approach that involves graph based methods for seed marking in image segmentation in the pre-processing phase to separate the hand from it's background, further it is fed as the input to ensemble models to predict and classify the gesture being performed, while allowing the users to be in their natural conditions.

Index Terms— image segmentation, graph algorithms, gesture classification, graph-cut, ensemble model, sign language, seed points selection, multi-threading

Sandesh BJ
PES University
Anirudh Srinivasa Raghavan
PES University
Akash Hamirwasia
PES University
Sneha Jain Ashok
PES University
Raghavendra Rao Suresh
PES University

[Download]

Cite: Sandesh B J, Anirudh Srinivasa Raghavan, Akash Hamirwasia, Sneha Jain Ashok, Raghavendra Rao Suresh, " Gesture Recognition Using Graph Method, " WCSE 2022 Spring Event: 2022 9th International Conference on Industrial Engineering and Applications, pp. 517-526, Sanya, China, April 15-18, 2022.