Modified Type-2 Fuzzy Gesture Space Induced Physical Disorder Recognition

Abstract

Gestures of a person symbolizing analogous physical disorder are not always exclusive. The gestural features of a subject suffering from the same physical disorder exhibit wide deviations for different instances. In presence of two or more gestural features, the variation of the attributes together makes the problem of physical disorder recognition more convoluted. This fluctuation is the main source of uncertainty in the physical disorder recognition problem, which has been addressed here using Type-2 Fuzzy Sets. First a type-2 fuzzy gesture space is assembled with the background knowledge of gestural features of different subjects for different physical disorders. Second, the physical disorder of an unknown gestural expression is concluded based on the consensus of the measured gestural features with the fuzzy gesture space. A fusion of Interval and General Type-2 Fuzzy Sets has been used to construct the fuzzy gestural space. The Interval Type-2 Fuzzy Set involves primary membership functions for M gestural features obtained from N subjects, each having L instances of gestural expressions for a given physical disorder. Along with that, the General Type-2 Fuzzy Set also includes the secondary memberships for individual primary membership curve, which has been obtained here by formulating and solving an optimization problem. To accomplish this, we have used Artificial Bee Colony. The secondary membership of a given source, characterizing the reliability in its primary membership assignment, is resolved based on the amalgamated knowledge of all the subjects’ primary membership functions. The adopted modified Type-2 Fuzzy Set-based uncertainty management strategy using gender independent training and testing databases has resulted in a classification accuracy of 93.48%.