Discriminative out-of-distribution detection for semantic segmentation

TL;DR

This paper introduces a discriminative method for detecting out-of-distribution pixels in semantic segmentation, training a dedicated OOD model to distinguish primary training data from a broader background dataset, improving OOD detection in high-resolution natural images.

Contribution

The paper proposes a novel OOD detection approach that trains a separate model to differentiate training data from a large background dataset, avoiding decisions based solely on the primary model's training data.

Findings

Outperforms previous OOD detection methods significantly.

Effectively identifies OOD pixels in high-resolution natural images.

Demonstrates robustness on WildDash dataset with out-of-distribution images.

Abstract

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.