Stochastic-YOLO: Efficient Probabilistic Object Detection under Dataset Shifts
This work addresses uncertainty estimation for object detection, which is important for applications requiring robustness to dataset shifts, but it is incremental as it builds on existing YOLO and MC-Dropout methods.
The paper tackles the challenge of evaluating uncertainty in object detection under dataset shifts by adapting YOLOv3 with Monte Carlo Dropout to generate uncertainty estimations, resulting in improved spatial uncertainties.
In image classification tasks, the evaluation of models' robustness to increased dataset shifts with a probabilistic framework is very well studied. However, object detection (OD) tasks pose other challenges for uncertainty estimation and evaluation. For example, one needs to evaluate both the quality of the label uncertainty (i.e., what?) and spatial uncertainty (i.e., where?) for a given bounding box, but that evaluation cannot be performed with more traditional average precision metrics (e.g., mAP). In this paper, we adapt the well-established YOLOv3 architecture to generate uncertainty estimations by introducing stochasticity in the form of Monte Carlo Dropout (MC-Drop), and evaluate it across different levels of dataset shift. We call this novel architecture Stochastic-YOLO, and provide an efficient implementation to effectively reduce the burden of the MC-Drop sampling mechanism at inference time. Finally, we provide some sensitivity analyses, while arguing that Stochastic-YOLO is a sound approach that improves different components of uncertainty estimations, in particular spatial uncertainties.