Fusion-driven deep feature network for enhanced object detection and tracking in video surveillance systems

Deepak Kumar Jain, Xudong Zhao, Chenquan Gan, Piyush Kumar Shukla, Amar Jain, Sourabh Sharma

Abstract: Highlights•Introduce an enhanced deep fused learning (EDFL) model for optimization.•Use a Dense Convolution Feature Fusion Network (D-ConvFFN) to enhance the feature fusion model.•Develop an object detection module called Enhanced RefineDet-based Fire Hawk (ERFH).•Assess the performance of the EDFL model using benchmark datasets.