{"slug": "c-gap-class-aware-and-online-prompting-improves-vision-language-models-on", "title": "C-GAP: Class-Aware and Online Prompting Improves Vision-Language Models on Imbalanced Classes", "summary": "Researchers introduced C-GAP, a framework that iteratively refines language prompts for frozen open-vocabulary detectors to improve detection of rare object classes without retraining or additional annotations. The method achieved up to 53% improvement in minority-class average precision and an 81% relative gain on COCO, demonstrating that composite captions combining scene descriptions and class-quantity context are critical for effective refinement.", "body_md": "arXiv:2607.09008v1 Announce Type: new\nAbstract: Safety-critical perception systems must reliably detect rare object classes within small label spaces, a setting that long-tailed detection methods, designed for hundreds of classes with dense annotation, fundamentally do not address. Open-vocabulary detectors offer a promising alternative, as they use natural language queries at inference time, making prompt quality a first-class lever for detection performance. We exploit this property to address class imbalance: rather than retraining models or collecting additional annotations, we ask whether iteratively refining the language prompts, fed to frozen detectors, can improve minority class detection. We introduce C-GAP Caption-Guided Augmentation and Prompting), a detector-agnostic, annotation-free framework that operates in two phases. First, we establish a composite caption baseline combining per-image scene descriptions with class-quantity context, which we show outperforms scene-description only or class-quantity-only prompts across multiple open-vocabulary architectures and benchmarks. Second, an LLM iteratively refines each image's caption individually, with trials triaged into accept, tentative, or regenerate buckets based on minority-class AP@0.5 against a dynamic threshold derived from the composite baseline. Refinement terminates early once sufficient AP@0.5 gain is achieved. No detector weights are updated at any stage. Our experiments shows that C-GAP improves minority-class average precision up to 53% over the baselines. On COCO, C-GAP improves minority-class AP@0.5 by ~81% relative over the composite baseline (17.69 -> 32.09). Experiments confirm that composite captions provide the critical foundation for effective refinement: using scene-description-only or class-quantity-only prompts as the refinement starting point yields diminishing returns, supporting both stages of C-GAP as necessary contributions.", "url": "https://wpnews.pro/news/c-gap-class-aware-and-online-prompting-improves-vision-language-models-on", "canonical_source": "https://arxiv.org/abs/2607.09008", "published_at": "2026-07-13 04:00:00+00:00", "updated_at": "2026-07-13 04:24:59.352079+00:00", "lang": "en", "topics": ["computer-vision", "natural-language-processing", "large-language-models", "artificial-intelligence"], "entities": ["C-GAP", "COCO"], "alternates": {"html": "https://wpnews.pro/news/c-gap-class-aware-and-online-prompting-improves-vision-language-models-on", "markdown": "https://wpnews.pro/news/c-gap-class-aware-and-online-prompting-improves-vision-language-models-on.md", "text": "https://wpnews.pro/news/c-gap-class-aware-and-online-prompting-improves-vision-language-models-on.txt", "jsonld": "https://wpnews.pro/news/c-gap-class-aware-and-online-prompting-improves-vision-language-models-on.jsonld"}}