Home

Awesome

Image Captioning and Text-to-Image Synthesis with textual data augmentation

This code run well under python2.7 and TensorFlow 0.11, if you use higher version of TensorFlow you may need to update the tensorlayer folder from TensorLayer Lib.

Usage

1. Prepare MSCOCO data and Inception model

2. Train image captioning model

3. Setup your paths

4. Train text-to-image synthesis with image captioning

Results

1. Here are some results on MSCOCO

<div align="center"> <img src="img/qualitative.jpeg" width="90%" height="30%"/> </div> <div align="center"> <img src="img/result.jpeg" width="80%" height="30%"/> </div>

2. Transfer learning on MHP dataset

<div align="center"> <img src="img/transferlearning.jpeg" width="100%" height="30%"/> </div>

Citation

@article{hao2017im2txt2im,
  title={I2T2I: LEARNING TEXT TO IMAGE SYNTHESIS WITH TEXTUAL DATA AUGMENTATION},
  author={Hao Dong, Jingqing Zhang, Douglas McIlwraith, Yike Guo},
  journal={ICIP},
  year={2017}
}