Under Convnet

Implementation of the paper "Adversarial Manipulation of Deep Representations". The code can manipulate the representation of an image in a deep neural network (DNN) to mimic those of other natural images, with only minor, imperceptible perturbations to the original image.

Dependencies

This code is written in python. To use it you will need:

• Caffe

In utils.py set caffe_root to your installation path.

Reference

If you found this code useful, please cite the following paper:

@article{sabour2015adversarial,
    title={Adversarial Manipulation of Deep Representations},
    author={Sabour, Sara and Cao, Yanshuai and Faghri, Fartash and Fleet,
        David J},
    journal={arXiv preprint arXiv:1511.05122},
    year={2015}
}

License

Apache License 2.0