Awesome

Code package that implements the image segmentation measures and reproduces all results from the papers:

Supervised Evaluation of Image Segmentation and Object Proposal Techniques<br/> Jordi Pont-Tuset and Ferran Marques, TPAMI 2015.

Measures and Meta-Measures for the Supervised Evaluation of Image Segmentation<br/> Jordi Pont-Tuset and Ferran Marques, CVPR 2013.

The project page contains updated evaluation, and browsable results from all techniques.

Version history:

v3.0: Generalization to Pascal Context, Pascal VOC, SBD.

v2.1: Bug fixes over 2.0. Stable release.

v2.0: Beta release that includes all experiments on the PAMI paper. It includes the following improvements:

• Easily evaluate any type of format via input converters.
• Contour detectors can also be included in the precision/recall curves.
• Updated list of state-of-the-art pre-computed results (up to ICCV 2015).
• Removed dependency from BSR.

v1.1: Stable release, bug fixes over v1.0, still at the CVPR level.

v1.0: Original beta release of the code, reproducing results of the CVPR paper.