Probabilistic 3d regression with projected huber distribution
This work addresses uncertainty estimation in 3D regression for computer vision applications, but it appears incremental as it builds on existing methods with specific property constraints.
The paper tackled the problem of estimating probability distributions for object localization from camera data by proposing a method that adheres to specific properties, resulting in uncertainties that correlate well with empirical errors and a mode that outperforms regression baselines.
Estimating probability distributions which describe where an object is likely to be from camera data is a task with many applications. In this work we describe properties which we argue such methods should conform to. We also design a method which conform to these properties. In our experiments we show that our method produces uncertainties which correlate well with empirical errors. We also show that the mode of the predicted distribution outperform our regression baselines. The code for our implementation is available online.