This repository contains the pytorch codes, trained models, and the new benchmarks described in our ACM MM 2019 paper "A New Benchmark and Approach for Fine-grained Cross-media Retrieval".

For more details, please visit our project page.

Results

• The MAP scores of bi-modality fine-grained cross-media retrieval of our FGCrossNet

• The MAP scores of multi-modality fine-grained cross-media retrieval of our FGCrossNet

Requirement

• pytorch, tested on [v1.0]
• CUDA, tested on v9.0
• Language: Python 3.6

Please visit our project page.

The trained models of our FGCrossNet framework can be downloaded from OneDrive, Google Drive or Baidu Cloud.

python audio.py

sh train.sh

sh test.sh

@inproceedings{he2019fine,
    Author = {Xiangteng He, Yuxin Peng, Liu Xie},
    Title = {A New Benchmark and Approach for Fine-grained Cross-media Retrieval},
    Booktitle = {Proc. of ACM International Conference on Multimedia (ACM MM)},
    Year = {2019}
} 

For any questions, feel free to open an issue or contact us. ([email protected])