Our research in image mining focusses on semantically improving common imaging applications. This page gives an overview of the different applications and offers downloads of source code and datasets.
Semantic Image Enhancement
Given an input image and a keyword, in this example strawberry, re-render the image in order to strengthen the semantic context.
Download: Matlab code + database (40MB)
Given an arbitrary color name in any language, estimate the corresponding sRGB values. This example visualizes periwinkle where the resulting sRGB values are: 169, 167, 212.
Browse more color names in more languages here: colorthesaurus.com
Download: Matlab code + database (60MB)
Given an arbitrary semantic expression, find a harmonic color palette that represents the semantic concept. the figure below reproduces automatically estimated color palettes for different keywords and harmonies.
Search your own color palettes here: www.koloro.org