NAU-QMUL: Utilizing BERT and CLIP for Multi-modal AI-Generated Image Detection
Fuente:
arXiv
Guardado en:
| Autores principales: | , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914357446705152 |
|---|---|
| author | Guo, Xiaoyu Zubiaga, Arkaitz |
| author_facet | Guo, Xiaoyu Zubiaga, Arkaitz |
| contents | With the aim of detecting AI-generated images and identifying the specific models responsible for their generation, we propose a multi-modal multi-task model. The model leverages pre-trained BERT and CLIP Vision encoders for text and image feature extraction, respectively, and employs cross-modal feature fusion with a tailored multi-task loss function. Additionally, a pseudo-labeling-based data augmentation strategy was utilized to expand the training dataset with high-confidence samples. The model achieved fifth place in both Tasks A and B of the `CT2: AI-Generated Image Detection' competition, with F1 scores of 83.16\% and 48.88\%, respectively. These findings highlight the effectiveness of the proposed architecture and its potential for advancing AI-generated content detection in real-world scenarios. The source code for our method is published on https://github.com/xxxxxxxxy/AIGeneratedImageDetection. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_23863 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | NAU-QMUL: Utilizing BERT and CLIP for Multi-modal AI-Generated Image Detection Guo, Xiaoyu Zubiaga, Arkaitz Computer Vision and Pattern Recognition Computation and Language With the aim of detecting AI-generated images and identifying the specific models responsible for their generation, we propose a multi-modal multi-task model. The model leverages pre-trained BERT and CLIP Vision encoders for text and image feature extraction, respectively, and employs cross-modal feature fusion with a tailored multi-task loss function. Additionally, a pseudo-labeling-based data augmentation strategy was utilized to expand the training dataset with high-confidence samples. The model achieved fifth place in both Tasks A and B of the `CT2: AI-Generated Image Detection' competition, with F1 scores of 83.16\% and 48.88\%, respectively. These findings highlight the effectiveness of the proposed architecture and its potential for advancing AI-generated content detection in real-world scenarios. The source code for our method is published on https://github.com/xxxxxxxxy/AIGeneratedImageDetection. |
| title | NAU-QMUL: Utilizing BERT and CLIP for Multi-modal AI-Generated Image Detection |
| topic | Computer Vision and Pattern Recognition Computation and Language |
| url | https://arxiv.org/abs/2602.23863 |