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

TL;DR

This paper introduces a differentiable polygon IoU loss for viewpoint-agnostic monocular 3D vehicle detection, improving convergence and accuracy by better handling projected 3D bounding box footprints.

Contribution

It proposes a novel, efficient algorithm for calculating IoU between convex polygons, enhancing 3D detection models' training with a new loss function.

Findings

PIoU loss converges faster than L1 loss.

Combining PIoU with L1 improves detection accuracy.

Enhanced performance on multiple state-of-the-art models.

Abstract

Monocular 3D object detection is a challenging task because depth information is difficult to obtain from 2D images. A subset of viewpoint-agnostic monocular 3D detection methods also do not explicitly leverage scene homography or geometry during training, meaning that a model trained thusly can detect objects in images from arbitrary viewpoints. Such works predict the projections of the 3D bounding boxes on the image plane to estimate the location of the 3D boxes, but these projections are not rectangular so the calculation of IoU between these projected polygons is not straightforward. This work proposes an efficient, fully differentiable algorithm for the calculation of IoU between two convex polygons, which can be utilized to compute the IoU between two 3D bounding box footprints viewed from an arbitrary angle. We test the performance of the proposed polygon IoU loss (PIoU loss) on three state-of-the-art viewpoint-agnostic 3D detection models. Experiments demonstrate that the proposed PIoU loss converges faster than L1 loss and that in 3D detection models, a combination of PIoU loss and L1 loss gives better results than L1 loss alone (+1.64% AP70 for MonoCon on cars, +0.18% AP70 for RTM3D on cars, and +0.83%/+2.46% AP50/AP25 for MonoRCNN on cyclists).