SAFE: Sensitivity-Aware Features for Out-of-Distribution Object Detection

TL;DR

This paper introduces SAFE, a novel feature extraction method using residual convolutional layers with batch normalization, which significantly improves out-of-distribution object detection without retraining the base detector.

Contribution

SAFE leverages residual convolutional layers with batch normalization to produce features that enhance OOD detection, avoiding retraining or generative models.

Findings

SAFE outperforms state-of-the-art OOD detectors on multiple benchmarks.

SAFE reduces FPR95 by 30.6% on OpenImages dataset.

SAFE does not require retraining the base detector.

Abstract

We address the problem of out-of-distribution (OOD) detection for the task of object detection. We show that residual convolutional layers with batch normalisation produce Sensitivity-Aware FEatures (SAFE) that are consistently powerful for distinguishing in-distribution from out-of-distribution detections. We extract SAFE vectors for every detected object, and train a multilayer perceptron on the surrogate task of distinguishing adversarially perturbed from clean in-distribution examples. This circumvents the need for realistic OOD training data, computationally expensive generative models, or retraining of the base object detector. SAFE outperforms the state-of-the-art OOD object detectors on multiple benchmarks by large margins, e.g. reducing the FPR95 by an absolute 30.6% from 48.3% to 17.7% on the OpenImages dataset.