Out-of-Distribution Detection for Monocular Depth Estimation
This addresses the issue of uncertainty due to lack of knowledge in depth estimation for applications like autonomous driving, though it is incremental as it builds on anomaly detection techniques.
The paper tackles the problem of detecting out-of-distribution (OOD) data in monocular depth estimation by using reconstruction error from an image decoder, achieving strong performance that outperforms existing uncertainty estimation methods on standard benchmarks like NYU Depth V2 and KITTI.
In monocular depth estimation, uncertainty estimation approaches mainly target the data uncertainty introduced by image noise. In contrast to prior work, we address the uncertainty due to lack of knowledge, which is relevant for the detection of data not represented by the training distribution, the so-called out-of-distribution (OOD) data. Motivated by anomaly detection, we propose to detect OOD images from an encoder-decoder depth estimation model based on the reconstruction error. Given the features extracted with the fixed depth encoder, we train an image decoder for image reconstruction using only in-distribution data. Consequently, OOD images result in a high reconstruction error, which we use to distinguish between in- and out-of-distribution samples. We built our experiments on the standard NYU Depth V2 and KITTI benchmarks as in-distribution data. Our post hoc method performs astonishingly well on different models and outperforms existing uncertainty estimation approaches without modifying the trained encoder-decoder depth estimation model.