An object detection method integrating global dependencies

The main task of object detection is to find the target object in the image and determine its class. The accuracy of the model in object detection can be enhanced by incorporating global dependencies when analyzing photos. However, constructing global dependencies is a significant challenge for convolutional neural networks. This paper presents a proposed solution for addressing the aforementioned issue, which involves integrating global dependencies into object detection. The network fuses the global dependencies into the feature map through the proposed global dependencies fusion block, which helps the model achieve better performance. At the same time, to ensure the model's performance, we propose a global dependencies concatenate block to enable the model to learn richer features. In conclusion, a series of experiments were performed on the PASCAL VOC dataset, resulting in the model achieving a mean average precision of 88.4%. Compared to the accuracy of other object detection methods, our model showed significant enhancements.