Computational Information Geometry Wonderland

My colleagues Paolo Piro and Michel Barlaud participated to the International Workshop on Image Analysis for Multimedia Interactive Services (WIAMIS), held in beautiful Italy in 2010:
We presented there some work on object classification using a weighted k-nn classification rule (k-nearest neighbours). The weights of labelled samples are automatically learnt using a boosting scheme.

For details, see report on arxiv: Boosting k-NN for categorization of natural scenes

Our algorithm bears the name MLNN for multiclass leveraged nearest neighbours. Frank.