AI RESEARCH
Does it Really Count? Assessing Semantic Grounding in Text-Guided Class-Agnostic Counting
arXiv CS.CV
•
ArXi:2605.02752v1 Announce Type: new Open-world text-guided class-agnostic counting (CAC) has emerged as a flexible paradigm for counting arbitrary object classes by using natural language prompts. However, current evaluation protocols primarily focus on standard counting errors within single-category images, overlooking a fundamental requirement: the ability to correctly ground the textual prompt in the visual scene.