Abstract
Human behaviour recognition systems hold the possibility of performing a variety of important assistive and management tasks in the development of ambient intelligent environments. However, the traditional non-fuzzy approaches for behaviour recognition using machine vision mostly rely on the assumptions such as known spatial locations and temporal segmentations or indispensably employ computationally expensive approaches such as sliding window search through a spatio-temporal volume. Hence, it is difficult for such traditional non-fuzzy methods to scale up the intelligent environments and handle the high-level of uncertainties available in real-world applications. To address these problems, this paper presents a system which is based on Interval Type-2 Fuzzy Logic Systems (IT2FLSs) whose parameters are optimized by the Big Bang-Big Crunch (BB-BC) algorithm which allows for robust behaviour recognition using machine vision in intelligent environments. We will present several experiments which were performed on the publicly available Weizmann human action dataset to fairly compare with the state-of-the-art algorithms. The experimental results demonstrate that the proposed optimization paradigm is effective in tuning the parameters of the membership functions and the rule base of the IT2FLSs to improve the recognition accuracy where the proposed IT2FLSs outperformed the Type-1 FLSs (T1FLSs) counterpart as well as outperforming other traditional non-fuzzy systems.