Finding Categorically Similar Images
Computers can categorise images incredibly accurately. For example, my house is categorised as a small wooden cottage. Can we use this information to find other similar houses? Google returns 11.2 billion images of houses; it is impossible to compare them all. If each comparison takes a millisecond, comparing them all will take 129 days. To overcome this, search algorithms reject dissimilar images without comparison, not always perfectly. We find that comparing the categorisation of images with measures of similarity derived from information theory can return higher quality results with fewer comparisons.
Keywords
Similarity Search, Databases, Information Theory, Artificial Intelligence
Staff
[Alan Dearle]{al}, [Richard Connor]{rchc}, Ben Claydon