Research map / Vision-language and multimodal / Multimodal LLMs
Fuzzy Contrastive Decoding to Alleviate Object Hallucination in Large Vision-Language Models
Jieun Kim, Jinmyeong Kim, Yoonji Kim, Sung-Bae Cho
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 FuzzyCD decoding | LLaVA-1.5 · POPE MSCOCO Random | accuracy (%): 82.90 -> 88.97 | Improved |
| add FuzzyCD decoding | LLaVA-1.5 · POPE MSCOCO Popular | accuracy (%): 81.00 -> 86.10 | Improved |
| add FuzzyCD decoding | LLaVA-1.5 · POPE MSCOCO Adversarial | accuracy (%): 78.60 -> 83.93 | Improved |
| add FuzzyCD decoding | InstructBLIP · POPE MSCOCO Random | accuracy (%): 80.63 -> 86.66 | Improved |
| add FuzzyCD decoding | LLaVA-1.5 · ROPE multi-object homogeneous | accuracy (%): 31.88 -> 46.44 | Improved |
| add FuzzyCD decoding | LLaVA-1.5 · ROPE multi-object wild | accuracy (%): 13.96 -> 21.12 | Improved |
| add FuzzyCD decoding | LLaVA-1.5 · ROPE multi-object heterogeneous | accuracy (%): 3.98 -> 7.45 | Improved |
| add FuzzyCD decoding | LLaVA-1.5 · MME total score | score: 556.34 -> 641.66 | Improved |
| add FuzzyCD decoding versus VDD on color | LLaVA-1.5 · MME color | score: 165 -> 160 | Hurt |
| add FuzzyCD decoding | LLaVA-1.5 · MME object count | score: 106.67 -> 163.33 | Improved |
| scale model from 7B to larger VLMs (ablation) | FuzzyCD · POPE | accuracy gain: 5.12%p LLaVA-1.5 -> 0.72%p QwenVL(32B) | Mixed |
| use Sobel instead of median image filter (ablation) | FuzzyCD · POPE | accuracy (%): 86.53 -> 89.33 | Improved |
Closest papers by meaning
- Mitigating Object Hallucination in Large Vision-Language Models via False Discovery Controlled Visual Data Splitting (NeurIPS 2026 · similarity 0.93)
- SECOND: Mitigating Perceptual Hallucination in Vision-Language Models via Selective and Contrastive Decoding (ICML 2025 · similarity 0.93)
- Mitigating Object Hallucinations in Large Vision-Language Models through Visual Contrastive Decoding (CVPR 2024 · similarity 0.93)
- Hallucination-aware Intermediate Representation Editing in Large Vision-Lanugage Models (ICLR 2026 · similarity 0.92)
- Reducing Hallucinations in Large Vision-Language Models via Latent Space Steering (ICLR 2025 · similarity 0.92)
- Alleviating Hallucinations in Large Vision-Language Models through Hallucination-Induced Optimization (NeurIPS 2024 · similarity 0.92)
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.
