EAMB-Net

This is the official code for the EAMBNet (TIM2024).

📎 Paper Link

• Article title: Image Aesthetics Assessment with Emotion-Aware Multi-Branch Network. Link
• Authors: Hangwei Chen, Feng Shao, Baoyang Mu, Qiuping Jinag
• Institution: Faculty of Electrical Engineering and Computer Science, Ningbo University

📃 Dependencies

• pytorch
• torchvision
• tqdm

📁 Train

We provide codes (ResNet50 version) that support AVA and PARA datasets.

• 1 Please download the pretrained emotion-aware model, and then put it in 'pretrain_emotion_model' folder
• 2 Please run 'main.py'.

⏬ Download

• You can download the AVA database at here: AVA
• You can download the PARA database at here: PARA
• You can download the pretrained emotion-aware model at Baidu Cloud. (password: eamb) or Google Drive

✨ Statement

This project is for research purpose only, please contact us for the licence of commercial use. For any other questions please contact 1010075746@qq.com or shaofeng@nbu.edu.cn

🔍 Citation

If our datasets and criteria are helpful, please consider citing the following papers.

@article{EAMBNet,
  title={Image Aesthetics Assessment with Emotion-Aware Multi-Branch Network},
  author={Hangwei Chen, Feng Shao, Baoyang Mu, Qiuping Jinag},
  journal={IEEE TIM},
  year={2024},
}