Deep_SBIR_tf

Implementation of Sketch Me That Shoe and Deep Spatial-Semantic Attention for Fine-grained Sketch-based Image Retrieval

And if you use the code for your research, please cite our paper:

@inproceedings{yu2016sketch,
        title={Sketch me that shoe},
        author={Yu, Qian and Liu, Feng and Song, Yi-Zhe and Xiang, Tao and Hospedales, Timothy M and Loy, Chen Change},
        booktitle={Computer Vision and Pattern Recognition (CVPR), 2016 IEEE Conference on},
        pages={799--807},
        year={2016},
        organization={IEEE}
}

@article{yu2017sketch,
        title={Sketch-a-net: A deep neural network that beats humans},
        author={Yu, Qian and Yang, Yongxin and Liu, Feng and Song, Yi-Zhe and Xiang, Tao and Hospedales, Timothy M},
        journal={International Journal of Computer Vision},
        volume={122},
        number={3},
        pages={411--425},
        year={2017},
        publisher={Springer}
}

Note:

(1) Please download the pre-trained model and well-trained models via https://drive.google.com/open?id=1249Emu4ol3TuF99EtWZohMEW4LNMeosp;

(2) Implementation by Caffe can be found here: https://github.com/seuliufeng/DeepSBIR.