I2 t
0 likes145 views
This document discusses image parsing to text generation. It describes how an Image to Text (I2T) system works by 1) detecting objects in an image using an image parsing model, 2) mapping detected objects to text labels using an ontology, and 3) generating a natural language description of the image based on the labels and their relationships. Examples are given of boat movements extracted from a video and described in text. References are provided on image parsing techniques, ontologies for mapping labels, and lexical databases used.
I2 t
- 1. Image Parsing to Text Generation 5310450025 |
- 2. I2T (Image to Text) I2T
- 4. 1) 3) ) 4)
- 6. And Or
- 8. And Or
- 11. LHI
- 15. dataset sport activity http://www.imageparsing.com
- 17. LHI
- 18. LHI label map
- 19. label- map Active Shape Models ASMs
- 22. PDF,XML OWL (Web Ontology Language) SPARQL + WordNet Text description
- 28. 1 I2T
- 30. Boat_1 enters the scene on water region at 19:50 Boat_2 follows Boat_1 between 35:36 and 37:32 Boat_2 turns right at 55:00
- 32. AOG
- 34. Yao, Benjamin Z and other. (2010). I2T: Image Parsing to Text Description. Cited January 14, 2013, Available from: http://ieeexplore.ieee.org/xpl/art icleDetails.jsp?tp=&arnumber=54 87377 McGuinness, Deborah L. and Harmelen, Frank van. (2012). OWL Web Ontology Language. Cited January 14, 2013, Available from: http://www.w3.org/TR/owl- features/ G. A. Miller. (2012). Wordnet: A lexical database for English. Cited January 14, 2013, Available from: http://wordnet.princeton.edu/