Multi-objective effective enhanced adaptive fusion technique using BAT algorithm for effective gait-based recognition

In biometric system, gait based recognition has more challenges as a result of changing appearance by several factors. To overcome these challenges, in our previous work, effective gait recognition using multi-objective enhanced adaptive fusion technique by BAT algorithm (EGRMEAFBAT) was proposed in which the most efficient features, most informative less efficient features and shape features from a gait sequence of an individual and gender identification were utilized to recognize the person. In this paper, we propose a novel fusion technique for efficient gait recognition named as efficient gait recognition using multi-objective effective enhanced adaptive fusion by BAT algorithm (EGRME²AFBAT). The proposed method is performed by considering additional features such as velocity moment and depth of both hands and legs along with features utilized in EGRMEAFBAT. Here, velocity moments and depth are measured by velocity measurement and feature vector calculation method respectively. Then, the extracted features are effectively fused using multi-objective effective enhanced adaptive fusion by BAT algorithm and the fused features are used for classification to recognize the person. The experimental results demonstrate that the proposed technique provides better accuracy compared to the gait recognition by EGRMEAFBAT.