Dropout Sampling for Robust Object Detection in Open-Set Conditions

TL;DR

This paper explores the application of Dropout Sampling for estimating label uncertainty in object detection, demonstrating improvements in detection performance under open-set conditions in robotic vision.

Contribution

It is the first to apply Dropout Sampling to object detection, showing how uncertainty can enhance detection in open-set scenarios.

Findings

12.3% increase in recall at same precision

15.1% increase in precision at same recall

Effective uncertainty estimation improves detection robustness

Abstract

Dropout Variational Inference, or Dropout Sampling, has been recently proposed as an approximation technique for Bayesian Deep Learning and evaluated for image classification and regression tasks. This paper investigates the utility of Dropout Sampling for object detection for the first time. We demonstrate how label uncertainty can be extracted from a state-of-the-art object detection system via Dropout Sampling. We evaluate this approach on a large synthetic dataset of 30,000 images, and a real-world dataset captured by a mobile robot in a versatile campus environment. We show that this uncertainty can be utilized to increase object detection performance under the open-set conditions that are typically encountered in robotic vision. A Dropout Sampling network is shown to achieve a 12.3% increase in recall (for the same precision score as a standard network) and a 15.1% increase in precision (for the same recall score as the standard network).