Multi-Stage CNN-Based Monocular 3D Vehicle Localization and Orientation Estimation

This paper aims to design a 3D object detection model from 2D images taken by\nmonocular cameras by combining the estimated bird's-eye view elevation map and\nthe deep representation of object features. The proposed model has a\npre-trained ResNet-50 network as its backend network and three more branches.\nThe model first builds a bird's-eye view elevation map to estimate the depth of\nthe object in the scene and by using that estimates the object's 3D bounding\nboxes. We have trained and evaluate it on two major datasets: a syntactic\ndataset and the KIITI dataset.\n