TPDS2024

This release corresponds to the code used for the experiments reported in our survey paper:

A State-of-the-Art Review with Code about Connected Components Labeling on GPUs
F. Bolelli, S. Allegretti, L. Lumetti, C. Grana
IEEE Transactions on Parallel and Distributed Systems.
BibTex, PDF

v3.4-alpha

This pre-release includes many state-of-the-art connected components labeling algorithms designed for GPU architectures. The complete list is available in the README of the project, additional algorithms wrt previous releases are: BRB, STAVA, RASMUSSON, ACCL, DLS, M8DLS, HA4, HA8.

TPAMI2021

This release includes multiple connected components algorithms generated by means of the graphgen framework, as described in:

Bolelli, F., Allegretti, S., & Grana, C. (2021). One DAG to rule them all. IEEE Transactions on Pattern Analysis and Machine Intelligence.

New algorithms are:

• Tagliatelle Labeling
• PRED++
• SAUF 3D
• SAUF++ 3D
• PRED 3D
• PRED++ 3D

IVPR2021

This release includes two new connected components algorithms specifically designed to label bitonal images i.e. images with 1 bit per pixel.

The algorithms (BRTS & BMRS) are detailed in the paper:

Lee, W., Allegretti, S., Bolelli, F., & Grana, C. (2021, August). Fast Run-Based Connected Components Labeling for Bitonal Images. In 2021 Joint 10th International Conference on Informatics, Electronics & Vision (ICIEV) and 2021 5th International Conference on Imaging, Vision & Pattern Recognition (icIVPR) (pp. 1-8). IEEE.

ICPR2020

Minor release that includes EPDT algorithms described in A Heuristic-Based Decision Tree for Connected Components Labeling of 3D Volumes.

TPDS2019

This version of the benchmark is described in "Optimized Block-Based Algorithms to Label Connected Components on GPUs". It includes the possibility of testing and evaluating GPU algorithms alongside CPU ones. Additional datasets and tests have been included to consider 3D labeling algorithms. New algorithms have been added to the benchmark.

JRTIP2018

This release introduces many improvements w.r.t. v1.0: additional tests, datasets, and algorithms have been implemented; algorithms are now templated on the label solver employed. A complete description of YACCLAB v2.0 can be found in "Towards reliable experiments on the performance of Connected Components Labeling algorithms".

ICPR2016

This is the first public release of the YACCLAB benchmark as described in "YACCLAB - Yet Another Connected Components Labeling Benchmark".