BROWSE
Detail
Conference Paper
Open Ended Object Detection: A Survey
Cited - time in 
Share 
Share
- Authors
- Seher Kanwal, Seung-Ik Lee
- Issue Date
- 2026-02
- Citation
- 한국로봇학회 종합 학술 대회 2026, pp.343-345
- Publisher
- 한국로봇학회
- Language
- English
- Type
- Conference Paper
- Abstract
- Humans naturally perceive their surroundings by identifying all objects but traditional object detectors lack this capability, as they operate within a predefined category set. Open Vocabulary detection attempts to relax this limitation by allowing recognition of novel objects through text prompts, yet it still depends on predefined category names at inference time. Open ended object detection takes a further step by identifying all objects and generating free labels without requiring any preset vocabulary. This survey paper provides a concise review of the recent approaches in open-ended object detection, highlights the strengths, and outline the challenges that must be addressed to build reliable detectors for real world applications.
- Keyword
- Open-ended, Generative, Object detection
- KSP Keywords
- Open-ended, Real-world applications, object detection