cs.CVNov 8, 2026

Partition, Refine, and Detect: Preserving Novel Category Capacity in Open-Vocabulary Object Detection

Authors: Wei Zhang, Zhu Teng, Zhixiang Chi, Junnan Zou, Yi Liu, Jianping Fan

Organizations: the Beijing Jiaotong University, Beijing 100044, China · University of Toronto, Toronto, ON M5S 2E8, Canada · Baidu Inc., Shenzhen, Guangdong 518000, China · institute of Intelligent Robotics and Advanced Manufacturing, Fudan University, Shanghai, China

Abstract

Open-vocabulary object detection (OVD) expects the model trained on base classes to recognize both base and novel classes at inference. Typical methods rely on CLIP-style vision-language embeddings, yet the learned space is often biased toward base categories, with limited effective capacity preserved for novel concepts. Such a base-dominated representation can compress novel semantics, leading to semantic collapse and mis-calibrated confidence at inference. To address this issue, we present PR-OVD, a partition-and-refine framework that explicitly reserves capacity for novel categories and strengthens category separation, while being seamlessly integrated into different detector architectures. Specifically, PR-OVD first performs coarse partitioning to build global divisions for base, novel, and an explicit background partition in both visual and textual modalities, preventing object-irrelevant embeddings from leaking into category-specific representations. It then applies category-level refinement within each partition to sharpen decision boundaries and alleviate the lack of fine-grained region-level pretraining in CLIP. To further provide reliable visual anchors for novel categories, we design a cognitively inspired Structure Prompt (SP) that encodes both egocentric attributes and allocentric context, facilitating stable partition-and-refine learning. Extensive experiments on two OVD benchmarks show that PR-OVD delivers an average 6.5% proportional improvement on novel categories over the state-of-the-art methods. Notably, PR-OVD also benefits other open-vocabulary tasks, including segmentation and classification, further confirming its effectiveness and generality.

Figures & tables

Explore similar work

CardsList
  1. NoOVD: Novel Category Discovery and Embedding for Open-Vocabulary Object Detection

    Mar 22, 2026Yupeng Zhang, Ruize Han, Zhiwei Chen +2Open-Vocabulary Object Detection

  2. DetRefiner: Model-Agnostic Detection Refinement with Feature Fusion Transformer

    May 11, 2026Soichiro Okazaki, Tatsuya Sasaki, Hiroki OhashiOpen-Vocabulary Object DetectionObject Detection

  3. Unveiling the Unknown: Open Vocabulary Object Detection with Scene Graphs

    Jun 4, 2026Yi Chen, Yinghao Lu, Zhehao Li +4Open-Vocabulary Object DetectionGraph Reasoning