ARTIFICIAL NEURAL NETWORK-BASED OBJECT DETECTION SYSTEM AND METHOD

摘要

An artificial neural network-based object detection system and method. Said method comprises: acquiring a three-dimensional point cloud, and acquiring a first feature map of the three-dimensional point cloud by using a trunk neural network (S101); using an attention branch neural network to process the first feature map, and acquiring a second feature map, each position of the second feature map comprising a predicted attention coefficient corresponding to the position, the second feature map being further used for acquiring a loss function of a target object, and the loss function being used for updating a network coefficient of the attention branch neural network (S102); and obtaining a prediction result according to the second feature map, the prediction result comprising position information of the target object (S103).