Beyond Known Objects: A Novel Framework for Open-Set Object Detection using Negative-Aware Norm

2026-05-04

AI Summary

This work addresses the challenge of simultaneously detecting both known and unknown objects in open-world scenarios by proposing the NAN-SPOT framework. Without retraining existing object detectors, NAN-SPOT requires only a few hundred images and minutes of lightweight fine-tuning to effectively estimate objectness through a novel Negative-Aware Norm metric introduced in hidden layers. The study reveals, for the first time, that pretrained detectors inherently possess the capacity to discriminate objectness, and establishes COCO-Open—the most comprehensively annotated benchmark for open-set object detection to date. The method significantly improves unknown object detection while preserving accuracy on known classes, outperforming existing approaches that rely on large-scale training in both efficiency and robustness.

Abstract

Open-Set Object Detection (OSOD) is crucial for autonomous driving, where perception systems must recognize and localize both known and previously unseen objects in complex, dynamic environments. While recent approaches deliver promising results, they often require retraining the detector extensively to learn objectness, which describes the likelihood that a bounding box tightly encloses a valid object, regardless of whether its category was learned during training. Deviating from existing work, we hypothesize that standard off-the-shelf detectors may already contain helpful cues for objectness, owing to their training on numerous and diverse known categories. Building on this idea, we propose NAN-SPOT, a training-light framework that does not require to retrain the base object detector and estimates objectness by leveraging a hidden layer metric called Negative-Aware Norm (NAN), requiring only minutes of training on just hundreds of images. To support comprehensive evaluation, we introduce COCO-Open, an expanded version of the existing COCO-Mixed dataset, increasing unknown object annotations from 433 to 1853, making it the most exhaustively labeled dataset for OSOD to the best of our knowledge. Experimental results demonstrate that NAN-SPOT achieves even better performance on unknown object detection than methods requiring heavy training, without compromising performance on known objects. This efficiency and robustness make NAN-SPOT a promising step towards open-world perception in autonomous driving.

Problem

• Open-Set Object Detection
• unknown objects
• objectness
• autonomous driving
• open-world perception

Innovation

• Open-Set Object Detection
• Negative-Aware Norm
• Training-Light Framework
• Objectness Estimation
• COCO-Open