Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2510.13080 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911316399095808 |
|---|---|
| author | Fu, Shuai Zhou, Jian Chen, Qi Jing, Huang Nguyen, Huy Anh Liu, Xiaohan Zeng, Zhixiong Ma, Lin Zhang, Quanshi Wu, Qi |
| author_facet | Fu, Shuai Zhou, Jian Chen, Qi Jing, Huang Nguyen, Huy Anh Liu, Xiaohan Zeng, Zhixiong Ma, Lin Zhang, Quanshi Wu, Qi |
| contents | Diffusion probabilistic models (DPMs) have demonstrated remarkable progress in generative tasks, such as image and video synthesis. However, they still often produce hallucinated samples (hallucinations) that conflict with real-world knowledge, such as generating an implausible duplicate cup floating beside another cup. Despite their prevalence, the lack of feasible methodologies for systematically quantifying such hallucinations hinders progress in addressing this challenge and obscures potential pathways for designing next-generation generative models under factual constraints. In this work, we bridge this gap by focusing on a specific form of hallucination, which we term counting hallucination, referring to the generation of an incorrect number of instances or structured objects, such as a hand image with six fingers, despite such patterns being absent from the training data. To this end, we construct a dataset suite CountHalluSet, with well-defined counting criteria, comprising ToyShape, SimObject, and RealHand. Using these datasets, we develop a standardized evaluation protocol for quantifying counting hallucinations, and systematically examine how different sampling conditions in DPMs, including solver type, ODE solver order, sampling steps, and initial noise, affect counting hallucination levels. Furthermore, we analyze their correlation with common evaluation metrics such as FID, revealing that this widely used image quality metric fails to capture counting hallucinations consistently. This work aims to take the first step toward systematically quantifying hallucinations in diffusion models and offer new insights into the investigation of hallucination phenomena in image generation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_13080 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Counting Hallucinations in Diffusion Models Fu, Shuai Zhou, Jian Chen, Qi Jing, Huang Nguyen, Huy Anh Liu, Xiaohan Zeng, Zhixiong Ma, Lin Zhang, Quanshi Wu, Qi Computer Vision and Pattern Recognition Diffusion probabilistic models (DPMs) have demonstrated remarkable progress in generative tasks, such as image and video synthesis. However, they still often produce hallucinated samples (hallucinations) that conflict with real-world knowledge, such as generating an implausible duplicate cup floating beside another cup. Despite their prevalence, the lack of feasible methodologies for systematically quantifying such hallucinations hinders progress in addressing this challenge and obscures potential pathways for designing next-generation generative models under factual constraints. In this work, we bridge this gap by focusing on a specific form of hallucination, which we term counting hallucination, referring to the generation of an incorrect number of instances or structured objects, such as a hand image with six fingers, despite such patterns being absent from the training data. To this end, we construct a dataset suite CountHalluSet, with well-defined counting criteria, comprising ToyShape, SimObject, and RealHand. Using these datasets, we develop a standardized evaluation protocol for quantifying counting hallucinations, and systematically examine how different sampling conditions in DPMs, including solver type, ODE solver order, sampling steps, and initial noise, affect counting hallucination levels. Furthermore, we analyze their correlation with common evaluation metrics such as FID, revealing that this widely used image quality metric fails to capture counting hallucinations consistently. This work aims to take the first step toward systematically quantifying hallucinations in diffusion models and offer new insights into the investigation of hallucination phenomena in image generation. |
| title | Counting Hallucinations in Diffusion Models |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2510.13080 |