This is a selection of my open source contributions.

Research code

VLFeat. The VLFeat open source library implements popular computer vision algorithms in a simple-to-use package with MATLAB bindings. It bundles algorithms such as SIFT, MSER, k-means, hierarchical k-means, kd-trees, agglomerative information bottleneck, quick shift.

VGG MKL. VGG MKL is an implementation of multiple kernel learning for image classification. An extended version of this code was used to implement one of the two best performing methods in the VOC 2009 PASCAL detection challenge.

VLPov. VLPov is a package to extract depth map and camera parameter from POV-Ray 3-D scenes.

Utility code

Autorights. Autorights is a companion script that can be used to automatically extracted formatted HTML documentation from the comments embedded in a set of M-Files. Text is automatically structured by using a symple and natural syntax in the comments.

Anaview. Anaview is a MATLAB function to generate anaglyphs from MATLAB figures.

Educational code

This page contains pointers to a number of educational implementations of intersting algorithms (e.g. SVM)

Obsolete code

Legacy SIFT and MSER implementations. SIFT and SIFT++ are respectively a MATLAB/C and command line/C++ implementation of the SIFT feature detector and descriptor. MSER is a MATLAB/C implementation of the MSER detector. All this is superseded by VLFeat.

Legacy bag-of-words code. Bag is a MATLAB implementation of a bag-of-feature algorithm for object category recognition.