Generating Images from Captions with Attention thumbnail
Pause
Mute
Subtitles not available
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Generating Images from Captions with Attention

Published on May 27, 20162584 Views

Motivated by the recent progress in generative models, we introduce a model that generates images from natural language descriptions. The proposed model iteratively draws patches on a canvas, while at

Related categories

Chapter list

Generating Images from Captions with Attention00:00
Motivation00:15
Novel Compositions00:52
General Idea01:11
Language Model (Bidirectional RNN)01:43
Image Model02:16
Model03:10
Alignment04:12
Sharpening05:00
Complete Model05:40
Main Dataset (Microsoft COCO)05:47
Flipping Colors06:37
Flipping Backgrounds06:59
Flipping Objects07:22
Examples of Alignment07:45
text2image <-> image2text08:01
More Results 08:44
Lower Bound of Log-Likelihood in Nats10:41
Qualitative Comparison11:12
Conclusions11:42
Thank You!12:19