Fake-HR1: Rethinking Reasoning of Vision Language Model for Synthetic Image Detection
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866913023647547392 |
|---|---|
| author | Jiang, Changjiang Sha, Xinkuan Yu, Fengchang Liu, Jingjing Liu, Jian Fang, Mingqi Zhang, Chenfeng Lu, Wei |
| author_facet | Jiang, Changjiang Sha, Xinkuan Yu, Fengchang Liu, Jingjing Liu, Jian Fang, Mingqi Zhang, Chenfeng Lu, Wei |
| contents | Recent studies have demonstrated that incorporating Chain-of-Thought (CoT) reasoning into the detection process can enhance a model's ability to detect synthetic images. However, excessively lengthy reasoning incurs substantial resource overhead, including token consumption and latency, which is particularly redundant when handling obviously generated forgeries. To address this issue, we propose Fake-HR1, a large-scale hybrid-reasoning model that, to the best of our knowledge, is the first to adaptively determine whether reasoning is necessary based on the characteristics of the generative detection task. To achieve this, we design a two-stage training framework: we first perform Hybrid Fine-Tuning (HFT) for cold-start initialization, followed by online reinforcement learning with Hybrid-Reasoning Grouped Policy Optimization (HGRPO) to implicitly learn when to select an appropriate reasoning mode. Experimental results show that Fake-HR1 adaptively performs reasoning across different types of queries, surpassing existing LLMs in both reasoning ability and generative detection performance, while significantly improving response efficiency. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_10042 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Fake-HR1: Rethinking Reasoning of Vision Language Model for Synthetic Image Detection Jiang, Changjiang Sha, Xinkuan Yu, Fengchang Liu, Jingjing Liu, Jian Fang, Mingqi Zhang, Chenfeng Lu, Wei Computer Vision and Pattern Recognition Artificial Intelligence Recent studies have demonstrated that incorporating Chain-of-Thought (CoT) reasoning into the detection process can enhance a model's ability to detect synthetic images. However, excessively lengthy reasoning incurs substantial resource overhead, including token consumption and latency, which is particularly redundant when handling obviously generated forgeries. To address this issue, we propose Fake-HR1, a large-scale hybrid-reasoning model that, to the best of our knowledge, is the first to adaptively determine whether reasoning is necessary based on the characteristics of the generative detection task. To achieve this, we design a two-stage training framework: we first perform Hybrid Fine-Tuning (HFT) for cold-start initialization, followed by online reinforcement learning with Hybrid-Reasoning Grouped Policy Optimization (HGRPO) to implicitly learn when to select an appropriate reasoning mode. Experimental results show that Fake-HR1 adaptively performs reasoning across different types of queries, surpassing existing LLMs in both reasoning ability and generative detection performance, while significantly improving response efficiency. |
| title | Fake-HR1: Rethinking Reasoning of Vision Language Model for Synthetic Image Detection |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2602.10042 |