Monocular 3D Object Detection in Cylindrical Images from Fisheye Cameras

by   Elad Plaut, et al.
General Motors

Detecting objects in 3D from a monocular camera has been successfully demonstrated using various methods based on convolutional neural networks. These methods have been demonstrated on rectilinear perspective images equivalent to being taken by a pinhole camera, whose geometry is explicitly or implicitly exploited. Such methods fail in images with alternative projections, such as those acquired by fisheye cameras, even when provided with a labeled training set of fisheye images and 3D bounding boxes. In this work, we show how to adapt existing 3D object detection methods to images from fisheye cameras, including in the case that no labeled fisheye data is available for training. We significantly outperform existing art on a benchmark of synthetic data, and we also experiment with an internal dataset of real fisheye images.


page 2

page 3

page 13

page 14


ParkingSticker: A Real-World Object Detection Dataset

We present a new and challenging object detection dataset, ParkingSticke...

Single-Stage Monocular 3D Object Detection with Virtual Cameras

While expensive LiDAR and stereo camera rigs have enabled the developmen...

Monocular 3D Vehicle Detection Using Uncalibrated Traffic Cameras through Homography

This paper proposes a method to extract the position and pose of vehicle...

Polygon Intersection-over-Union Loss for Viewpoint-Agnostic Monocular 3D Vehicle Detection

Monocular 3D object detection is a challenging task because depth inform...

CenterLoc3D: Monocular 3D Vehicle Localization Network for Roadside Surveillance Cameras

Monocular 3D vehicle localization is an important task in Intelligent Tr...

Camera Elevation Estimation from a Single Mountain Landscape Photograph

This work addresses the problem of camera elevation estimation from a si...

Please sign up or login with your details

Forgot password? Click here to reset