Crossing textual and visual content in different application scenarios thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Crossing textual and visual content in different application scenarios

Published on Feb 14, 20083216 Views

In this presentation, we present a method based on Trans-media Pseudo-Relevance Feedback that allows crossing visual and textual content through multimodal knowledge base. The main idea is to use one

Chapter list

Crossing textual and visual content in different application scenarios00:00
The main idea - 100:17
An example scenario - 101:05
An example scenario - 202:11
The main idea - 202:53
Outline02:59
Outline03:15
Image similarity03:16
Low-level features03:52
Visual vocabulary with a GMM04:16
The Fisher vector04:24
Similarity between images04:50
Outline05:11
Example of retrieved images in our TBAS05:13
Outline05:44
Image metadata examples in TBAS using GVC05:45
Outline06:38
Text Representation06:40
Outline07:13
Fusion between image and text07:21
Intermediate level fusion08:16
Pseudo feedback (PF)09:14
Outline09:46
Text illustration09:48
Image annotation11:00
Examples of auto-annotation from the repository11:36
Outline12:02
Information retrieval12:07
Retrieval results of ImageClefPhoto14:00
Outline14:51
Relating text and image through a repository15:01
Examples of a p es o text and images linked by the TBAS15:47
Conclusion16:20
Thank you16:55
Performance of our GVC20:12