Discriminative out-of-distribution detection for semantic segmentation
This addresses the issue of unavoidable failures in real-life computer vision applications due to incomplete visual ontologies, though it is incremental as it builds on existing OOD detection approaches.
The paper tackles the problem of out-of-distribution (OOD) detection in semantic segmentation by proposing a discriminative method that trains a dedicated OOD model using a primary training set and a large background dataset, achieving results that outperform previous work by a wide margin on the WildDash test dataset.
Most classification and segmentation datasets assume a closed-world scenario in which predictions are expressed as distribution over a predetermined set of visual classes. However, such assumption implies unavoidable and often unnoticeable failures in presence of out-of-distribution (OOD) input. These failures are bound to happen in most real-life applications since current visual ontologies are far from being comprehensive. We propose to address this issue by discriminative detection of OOD pixels in input data. Different from recent approaches, we avoid to bring any decisions by only observing the training dataset of the primary model trained to solve the desired computer vision task. Instead, we train a dedicated OOD model which discriminates the primary training set from a much larger "background" dataset which approximates the variety of the visual world. We perform our experiments on high resolution natural images in a dense prediction setup. We use several road driving datasets as our training distribution, while we approximate the background distribution with the ILSVRC dataset. We evaluate our approach on WildDash test, which is currently the only public test dataset that includes out-of-distribution images. The obtained results show that the proposed approach succeeds to identify out-of-distribution pixels while outperforming previous work by a wide margin.