research-papers

Interesting paper about how using weaker classifiers (kmeans) can get comparable results to more sophisticated ones like deep neural nets with the decisions made before training a model like statistical whitening, picking a large number of features, etc

Real time face detection method most cameras use today, discusses ensemble learning methods (adaboost) and a clever way to detect features in one O(n) pre-processing step

Discusses how you can reconstruct an objects 3d shape just by analyzing how light reflects off an object's surface

Accessible introduction to the machinery underneath ST and IO