Distance metric learning in Python

Algorithms

This python module implements two distance metric learning algorithms for learning metrics from pairwise similarity constraints:

• Pairwise constrained component analysis (PCCA) [1]
• Cross modal metric learning (CMML) [2]

Disclaimer

This code has been re-written from scratch and is mainly untested. Results obtained using this code are not guaranteed to match those published in [1] and [2]. Some preliminary tests seem to show results similar to those obtained in [1] on LFW.

Usage condition

Any publication made using this code or a modification, adaptation or traduction of this code should mention the publications [1] and [2]

Requirements

This module depends on the lgbopt module available here: https://github.com/alexis-mignon/python-lgbopt

References

[1] PCCA: A new approach for distance learning from sparse pairwise constraints. Alexis Mignon, Frédéric Jurie. Computer Vision and Pattern Recognition (CVPR) 2012.

[2] CMML: a New Metric Learning Approach for Cross Modal Matching. Alexis Mignon, Frédéric Jurie; Asian Conference on Computer Vision (ACCV) 2012.