A Fuzzy C Means Clustering Approach for Gesture Recognition in Healthcare
Monalisa Pal, Sriparna Saha, Amit Konar · 2014
The aim of this novel work is to recognize 12 health care linked gestures from young individuals of 20- 40 years of age group. Due to constant sitting in a specific posture for deskbound jobs, functioning of joints and muscles of persons are deteriorated. The scope of this work is to recognize the early stage symptoms of those physical disorders and notify the persons about their decaying health. This medical knowledge based system also prescribes an exercise based on recognized disorder after consulting doctors. The work deals with principal component analysis for linear dimensionality reduction and recognition using fuzzy c means algorithm. The overlapping of gestures in feature space demonstrates the fuzziness of the input. This easy but effective technique provides a high accuracy of 96.0201% in 0.0439 second. The results are compared with those obtained from other standard clustering methods using McNemar's Test, thereby validating the proposed method.