A Lightweight DualStream-BioInceptNet CNN for Real-Time Suspect Identification using Face and Gait Recognition

Biometric recognition systems are critical to the security and surveillance use case, but they often have challenges in the real world, especially with occluded or poorly lit faces. In such situations, gait-based recognition is a suitable alternative. This research was inspired by how humans identify others naturally through facially identified features or gait/gestures. Based on this natural composition of human recognition strategies, developed a dual modality model incorporating a face and gait recognition model. The proposed a lightweight CNN architecture called DualStream-BioInceptNet, which combines inception modules with compact dense layers to process each frame to extract spatial information in an effective way of individuals. The model was purposely designed to avoid computationally heavy or recurrent layers, which allows implementation at a level that would provide both efficiency and real-time capability, including technology with poor computing capabilities as needed in modern biometrics under constrained conditions. From evaluation of the model on the customised dataset with variable real-world conditions, which achieve a recognition accuracy of 97.92%. This demonstrates the ability of model to generalise between different scenarios whilst providing robustness, and evidencing real-world usage. The architecture demonstrated a robust equipoise among accuracy and computational efficiency, which is desirable for instantaneous deployment in environments with limited resources. Its design reflects the way humans recognize others, and is effective for practical biometric recognition tasks.