Fully Convolutional Network for Object Direction Estimation in the Wild
Min Du, Guoyou Wang · 2019
In recent years, object detection has been widely applied in various fields, such as aircraft navigation, and power line inspection. However, the object pose estimation in complex background still remains challenging. In this paper, an end-to-end object direction regression system based on the fully convolutional network is proposed. The system can learn deep features from data and use convolutional layers to regress the direction angle of the object. The experiments show that this method can accurately estimate the direction angle of the object from the image.