Calibrating Uncertainties in Object Localization Task
arXiv:1811.11210
Abstract
In many safety-critical applications such as autonomous driving and surgical robots, it is desirable to obtain prediction uncertainties from object detection modules to help support safe decision-making. Specifically, such modules need to estimate the probability of each predicted object in a given region and the confidence interval for its bounding box. While recent Bayesian deep learning methods provide a principled way to estimate this uncertainty, the estimates for the bounding boxes obtained using these methods are uncalibrated. In this paper, we address this problem for the single-object localization task by adapting an existing technique for calibrating regression models. We show, experimentally, that the resulting calibrated model obtains more reliable uncertainty estimates.
References in corpus (6)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
- On Calibration of Modern Neural Networks
- Accurate Uncertainties for Deep Learning Using Calibrated Regression
- Fast and Scalable Bayesian Deep Learning by Weight-Perturbation in Adam
- Robustly representing uncertainty in deep neural networks through sampling
Cited by in corpus (6)
- TransMorph: Transformer for unsupervised medical image registration
- Multivariate Confidence Calibration for Object Detection
- Parameters Estimation for the Cosmic Microwave Background with Bayesian Neural Networks
- Uncertainty for Identifying Open-Set Errors in Visual Object Detection
- Notes on the Behavior of MC Dropout
- Semi-Local 3D Lane Detection and Uncertainty Estimation