FocusOVD: Geometric and Contextual Feature Refinement for Open-Vocabulary Small-Object Detection
To address these challenges, we propose FocusOVD, an open-vocabulary object detector that enhances small-object representations without modifying the frozen VLM backbone.
ProofPaper ↗
Key points
- Open-vocabulary object detection aims to detect both seen and unseen categories by leveraging the semantic knowledge of vision-language models.
- However, existing methods often struggle with small objects due to insufficient preservation of high-resolution details, ineffective information aggregation across feature levels, and limited region-level discriminability.
- Extensive experiments on COCO-OVD and LVIS-OVD demonstrate the effectiveness of FocusOVD.
- On COCO-OVD, FocusOVD achieves a novel-category AP of 36.3%, outperforming CORA, UniDetector, and SAMP by 1.3%, 1.1%, and 1.5%, respectively; on LVIS-OVD, it achieves an overall AP of 33.7%, surpassing the strongest competing method by 0.6%.
Sources (1)
- [1]FocusOVD: Geometric and Contextual Feature Refinement for Open-Vocabulary Small-Object DetectionarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 20, 01:45 AM
To address these challenges, we propose FocusOVD, an open-vocabulary object detector that enhances small-object representations without modifying the frozen VLM backbone.
Open-vocabulary object detection aims to detect both seen and unseen categories by leveraging the semantic knowledge of vision-language models.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 20, 2026Qwen-Audio-3.1-ASR Technical Report
- Oct 20, 2026PAR: Progress-Aligned Recurrence for Length Generalization in State Tracking
- Oct 20, 2026DiffResearch: Accelerating Deep Research with Diffusion Language Models
- Oct 20, 2026Large Flow Language Models
- Oct 19, 2026Look Before You Reason: Implicit Visual Thinking for Efficient Multimodal Reasoning
- Oct 18, 2026Finite Interaction Lens: Structured Precision Sharing for Vision-Language Models