Attentive Semantic Video Generation using Captions
Tensorflow implementation for the paper Attentive Semantic Video Generation using Captions by Tanya Marwah*, Gaurav Mittal* and Vineeth N. Balasubramanian accepted at International Conference on Computer Vision 2017 (ICCV 2017) (*Equal Contribution).
![](https://user-images.githubusercontent.com/28288044/31856492-58956022-b690-11e7-9acd-31076e91d50a.jpg)
Proposed network architecture for attentive semantic video generation with captions.
![](https://camo.githubusercontent.com/33176e78817fc5d47ae3decb050d549a1edffcbedb8ecd3199b9eff45e66441e/687474703a2f2f692e696d6775722e636f6d2f477673627535372e676966) |
![](https://camo.githubusercontent.com/b1adb1dadc3fa4207eaf099f0ca224c909cf855fdf364fa55eaecfa43eae96a1/687474703a2f2f692e696d6775722e636f6d2f5561736c5763692e676966) |
digit 6 is moving up and down |
digit 3 is moving left and right |
![](https://user-images.githubusercontent.com/28288044/31858691-12f254d0-b6cc-11e7-9f67-1b8aaa457028.gif) |
person 4 is walking left to right |
Example of Spatio Temporal Style Transfer
![](https://camo.githubusercontent.com/57c9a6f3e8bbd49dd658469408a6ddb45abb30db0c5b46b9a2bb54c2eecfc40e/687474703a2f2f692e696d6775722e636f6d2f3051747950536a2e676966) |
Caption 1: digit 4 is moving up and down Caption 2: digit 4 is moving left and right |
![](https://camo.githubusercontent.com/bf28f4c17d4cd5e331d8655b3d2ab618e1497f852422889f19b23a44a69ad2ac/687474703a2f2f692e696d6775722e636f6d2f757361525461442e676966) |
![](https://camo.githubusercontent.com/feb377d2df233e080fc805294faf359bbc0dec7ae6b0f7db3d3c877561937f51/687474703a2f2f692e696d6775722e636f6d2f4a69554d3348592e676966) |
Caption 1: digit 4 is moving up and down Caption 2: digit 9 is moving left and right |
Caption 1: digit 5 is moving left and right Caption 2: digit 9 is moving up and down |
![](https://user-images.githubusercontent.com/28288044/31858692-12fd255e-b6cc-11e7-9da7-b85c1f0e54fc.gif) |
Caption 1: person 10 is walking left to right Caption 2: person 10 is walking right to left |