Object detection method based on dense connection and feature fusion

Object detection is a basic task in the field of computer vision and is widely used in various fields. However, there is also low detection performance caused by object scale changes and low feature extraction capabilities of the network, which makes the utilization of multi-scale features low. Therefore, this paper proposes a method based on dense connection and feature fusion. In this method, a dense connection module is designed to improve the ability of the network to extract features and improve the utilization of multi-scale features; a feature fusion module is designed to integrate feature information. In addition, The loss function uses Focal loss as classification loss and GIoU as positioning loss. The Tensorflow deep learning framework is used to deploy the network, and experiments are conducted on the VOC2007 and 2012 data sets to verify the effectiveness of the proposed method and compare it with the current method.