Generalized Object Detection on Fisheye Cameras for Autonomous Driving: Dataset, Representations and Baseline

TL;DR

This paper investigates improved object detection methods for fisheye cameras in autonomous driving, proposing novel representations and a new dataset to address distortion issues, leading to significant accuracy improvements.

Contribution

It introduces a new curved bounding box model, a curvature adaptive polygon sampling method, and a comprehensive dataset for fisheye camera object detection in autonomous driving.

Findings

Polygon model improves mIoU accuracy by 40.3%.

Curvature adaptive sampling increases mAP by 4.9%.

First detailed study on fisheye object detection for autonomous driving.

Abstract

Object detection is a comprehensively studied problem in autonomous driving. However, it has been relatively less explored in the case of fisheye cameras. The standard bounding box fails in fisheye cameras due to the strong radial distortion, particularly in the image's periphery. We explore better representations like oriented bounding box, ellipse, and generic polygon for object detection in fisheye images in this work. We use the IoU metric to compare these representations using accurate instance segmentation ground truth. We design a novel curved bounding box model that has optimal properties for fisheye distortion models. We also design a curvature adaptive perimeter sampling method for obtaining polygon vertices, improving relative mAP score by 4.9% compared to uniform sampling. Overall, the proposed polygon model improves mIoU relative accuracy by 40.3%. It is the first detailed study on object detection on fisheye cameras for autonomous driving scenarios to the best of our knowledge. The dataset comprising of 10,000 images along with all the object representations ground truth will be made public to encourage further research. We summarize our work in a short video with qualitative results at https://youtu.be/iLkOzvJpL-A.