This is the source code of our TIP 2018 paper "Modality-specific Cross-modal Similarity Measurement with Recurrent Attention Network", Please cite the following paper if you find our code useful.
Yuxin Peng, Jinwei Qi and Yuxin Yuan, “Modality-specific Cross-modal Similarity Measurement with Recurrent Attention Network”, IEEE Transactions on Image Processing (TIP), DOI:10.1109/TIP.2018.2852503, 2018. [arxiv]
Our code is based on torch. You need to first install torch as follows:
# in a terminal, run the commands WITHOUT sudo
git clone https://github.com/torch/distro.git ~/torch --recursive
cd ~/torch; bash install-deps;
./install.sh
See here for more details when you encounter problems during installation.
The code is tested on Ubuntu 14.04.5 LTS, Lua 5.1.
Data Preparation: we use wiki dataset as example, and the data should be put in ./i2t_attention/data/ and ./t2i_attention/data/.
The data files can be download from the link and unzipped to the above path.
run sh run.sh to train models and extract features, then run the following commands to calculate mAP:
cd ./calMAP
matlab -r "evalMAPMerge"
If you are interested in cross-media retrieval, you can check our recently published overview paper on IEEE TCSVT:
Yuxin Peng, Xin Huang, and Yunzhen Zhao, "An Overview of Cross-media Retrieval: Concepts, Methodologies, Benchmarks and Challenges", IEEE Transactions on Circuits and Systems for Video Technology (TCSVT), 2017.[PDF]
Welcome to our Benchmark Website and Laboratory Homepage for more information about our papers, source codes, and datasets.