Detection of behavioural pattern through human-gait image using soft-computing technique

Abstract

In the recent times, it has been observed that most of the biometrical research work has been carried out either using template-based or knowledge-based approach. In the present paper, behavioural patterns have been determined through human-gait image through knowledge-based modelling using soft-computing technique. Very little amount of work has been carried out with knowledge-based or model-based approach using soft-computing tools like artificial neural network, fuzzy set theory and genetic algorithm. In the present work, soft-computing technique has been applied for the detection of behavioural patterns like: normal behaviour and abnormal behaviour. The work has been carried out in two phases. In the first phase, a knowledge-based model called artificial noise-free human-gait model (ANFHGM) has been formed. The model has been formed by considering an input human-gait image that has been enhanced and compressed for the removal of distortion with loss-less information. Later on it has been segmented using the mechanism for the detection of the region of interest along with object of interest. Hence the relevant geometric features have been extracted and stored in the form of a corpus. In the second phase, ANFHGM has been utilised for the determination of behavioural patterns with a test human-gait image. The process has been done by considering an unknown human-gait image for the detection of normal and abnormal behaviour patterns of the subject. For this an algorithm, known as BPDSC (Behavioural pattern detection using soft-computing) has been proposed in the present paper. The algorithm has been tested with 15 subjects of varying ages on a flat and plain surface. The result has been found very satisfactory with the tested data sets.