About this document
Autonomous Image Model Building Techniques by saviclosmi is a document available to read on EtoBox.
The document summarizes research on autonomous visual model building based on image crawling through internet search engines. It discusses using generalized multiple instance learning (GMIL) and uncertain labeling density (ULD) to build semantic models from image data gathered from search engines without manual labeling. The approach involves using GMIL and ULD to learn the most informative examples from image bags and build models to represent visual concepts for indexing images.
- Author
- saviclosmi
- Language
- EN