Research map / Detection and segmentation / Open-set and few-shot detection
Referring to Any Person
Qing Jiang, Lin Wu, Zhaoyang Zeng, Tianhe Ren, Yuda Xiong, Yihao Chen, Liu Qin, Lei Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add rejection data to HumanRef training (ablation) | RexSeek-7B · HumanRef Benchmark (rejection subset) | rejection score: 0 -> 541 | Improved |
| load the checkpoint after stage 2 instead of stage 1 before HumanRef fine-tuning (ablation) | RexSeek-3B · HumanRef Benchmark (average) | DensityF1: 68.2 -> 72.2 | Improved |
| load the checkpoint after stage 3 (general multimodal SFT) instead of stage 2 (ablation) | RexSeek-3B · HumanRef Benchmark (average) | DensityF1: 72.2 -> 73.0 | Improved |
| load the checkpoint after stage 3 instead of stage 1 (ablation) | RexSeek-3B · HumanRef Benchmark (average) | recall: 73.9 -> 77.9 | Improved |
| use RexSeek instead of ChatRex-7B for zero-shot object referring | RexSeek-7B · RefCOCOg val | accuracy: 88.8 -> 84.0 | Hurt |
| use RexSeek instead of Qwen2.5-VL-7B | RexSeek-7B · HumanRef Benchmark (average) | DensityF1: 56.2 -> 82.3 | Improved |
| use RexSeek instead of ChatRex-7B | RexSeek-7B · HumanRef Benchmark (average) | DensityF1: 55.6 -> 82.3 | Improved |
| compare RexSeek with Molmo-7B-D (point-in-mask criterion) on rejection | RexSeek-7B · HumanRef Benchmark (rejection subset) | rejection score: 68.6 -> 54.1 | Hurt |
| use RexSeek instead of the all-person baseline | RexSeek-7B · HumanRef Benchmark (average) | DensityF1: 18.9 -> 82.3 | Improved |
Closest papers by meaning
- A Large-Scale Human-Centric Benchmark for Referring Expression Comprehension in the LMM Era (NeurIPS 2024 · similarity 0.87)
- Rex-Thinker: Grounded Object Referring via Chain-of-Thought Reasoning (ICLR 2026 · similarity 0.86)
- A Pedestrian is Worth One Prompt: Towards Language Guidance Person Re-Identification (CVPR 2024 · similarity 0.85)
- RefDetector: A Simple Yet Effective Matching-based Method for Referring Expression Comprehension (AAAI 2025 · similarity 0.84)
- InstructDET: Diversifying Referring Object Detection with Generalized Instructions (ICLR 2024 · similarity 0.84)
- OneRef: Unified One-tower Expression Grounding and Segmentation with Mask Referring Modeling (NeurIPS 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
