Interactive Neural Network for Object Detection in YOLOv5 and YOLOv8

Object detection is widely used for detecting any moving/stable objects. In order to develop the object detection techniques, YOLOv5 and YOLOv8 are being analyzed with Attention Network modules. The dataset is being created with the flow of real samples taken by overhead luggage of the bus. Dataset is being augmented and with a highly trained network, the detection results have been further developed with a new Convolution Network module. A new Attention Convolution Network is being created for better results in detection which is a lightweight architecture. It incorporates a Local and Global Interactive Module (LGIM), which directly affects the dependency by combining the local and global parts. At the same time, the Enhancing Channel and Space Interaction (ECSI) module is demonstrated to minimize unneeded background noise by enabling two components combination as High and Low Order Global Attention.