Detecting Out-of-distribution Objects Using Neuron Activation Patterns

TL;DR

This paper introduces NAPTRON, a novel method for detecting out-of-distribution objects in detection models, outperforming existing techniques without compromising in-distribution accuracy, and provides a comprehensive benchmark across scenarios and detectors.

Contribution

The paper presents NAPTRON, a new approach leveraging neuron activation patterns for OOD detection in object detection, and establishes the largest open-source benchmark for this task.

Findings

NAPTRON outperforms state-of-the-art OOD detection methods.

The approach maintains in-distribution detection performance.

Benchmark results across multiple scenarios and detectors are provided.

Abstract

Object detection is essential to many perception algorithms used in modern robotics applications. Unfortunately, the existing models share a tendency to assign high confidence scores for out-of-distribution (OOD) samples. Although OOD detection has been extensively studied in recent years by the computer vision (CV) community, most proposed solutions apply only to the image recognition task. Real-world applications such as perception in autonomous vehicles struggle with far more complex challenges than classification. In our work, we focus on the prevalent field of object detection, introducing Neuron Activation PaTteRns for out-of-distribution samples detection in Object detectioN (NAPTRON). Performed experiments show that our approach outperforms state-of-the-art methods, without the need to affect in-distribution (ID) performance. By evaluating the methods in two distinct OOD scenarios and three types of object detectors we have created the largest open-source benchmark for OOD object detection.