EW-DETR: Evolving World Object Detection via Incremental Low-Rank DEtection TRansformer

2026-02-24

AI Summary

This work addresses the challenges of object detection in real-world scenarios—specifically, class-incremental learning, domain shift, and unknown object recognition—under the strict constraint of no access to historical data. To this end, the authors propose EW-DETR, a novel framework that introduces the Evolving World Object Detection (EWOD) paradigm. Built upon the DETR architecture, EW-DETR integrates three key components: incremental LoRA adapters, a Query-Norm Objectness Adapter, and an Entropy-Aware Unknown Mixing module, which jointly enable robust detection without replay. The study also introduces FOGS, a comprehensive evaluation metric to holistically assess performance. Extensive experiments on the Pascal Series and Diverse Weather benchmarks demonstrate a 57.24% improvement in FOGS over existing methods, substantiating the framework’s effectiveness and generalization capability.

Technology Category

• Computer Vision: Object Detection & Categorization
• Machine Learning: Adversarial Learning & Robustness
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

Real-world object detection must operate in evolving environments where new classes emerge, domains shift, and unseen objects must be identified as"unknown": all without accessing prior data. We introduce Evolving World Object Detection (EWOD), a paradigm coupling incremental learning, domain adaptation, and unknown detection under exemplar-free constraints. To tackle EWOD, we propose EW-DETR framework that augments DETR-based detectors with three synergistic modules: Incremental LoRA Adapters for exemplar-free incremental learning under evolving domains; a Query-Norm Objectness Adapter that decouples objectness-aware features from DETR decoder queries; and Entropy-Aware Unknown Mixing for calibrated unknown detection. This framework generalises across DETR-based detectors, enabling state-of-the-art RF-DETR to operate effectively in evolving-world settings. We also introduce FOGS (Forgetting, Openness, Generalisation Score) to holistically evaluate performance across these dimensions. Extensive experiments on Pascal Series and Diverse Weather benchmarks show EW-DETR outperforms other methods, improving FOGS by 57.24%.

Problem

• Evolving World Object Detection
• Incremental Learning
• Domain Adaptation
• Unknown Detection
• Exemplar-Free

Innovation

• Incremental Learning
• Domain Adaptation
• Unknown Object Detection
• LoRA Adapters
• DETR