Segmentation Parameter Tuner (SPT)

Segmentation Parameter Tuning (SPT), is an open source software designed for automatic tuning of segmentation parameters based on a number of optimization algorithms using different quality metrics as fitness functions.


The current version of SPT is 3.9 which includes many features such as:

  1. Many raster formats supported (provided by GDAL).
  2. Metaheuristics available to find the best set of parameters:
    1. Single-solution (S-metaheuristics)
      1. Generalized Pattern Search (GPS)
      2. Mesh Adaptive Direct Search (MADS)
      3. Nelder-Mead (NM)
    2. Population-based (P-metaheuristics)
      1. Differential Evolution (DE)
    3. Hybrid metaheuristics
      1. DE + GPS
      2. DE + MADS
      3. DE + NM
  3. Segmentation algorithms available:
    1. Multiresolution Region Growing
    2. EDISON Meanshift
    3. Graph-based (Felzenszwalb’s Segmentation)
    4. SPRING
    5. MultiSeg (for Optical and Radar images).
  4. Metrics available:
    1. Hoover Index
    2. Area-Fit-Index
    3. F-measure
    4. Shape Index
    5. Reference Bounded Segments Booster
    6. Rand Index
    7. Segmentation Covering
  5. Reference tool: Create, Edit and Save segmentation references in ESRI Shapefile format.
  6. Assessment tool: Measure the quality of a segmentation outcome (raster image or shapefile) with respect to a certain reference.

Packages

SPT 3.9 (last update: September 1st, 2016)

SPT 2.2 (Based on Genetic Algorithms)

Documentation

SPT User Guide 3.1.0 (last update: June 7th)
SPT User Guide 2.2

Demos

SPT Demo 1 (TOLOMEO Project: Final Meeting)
SPT Demo 2 (XVII SBSR: Oral Presentation)

 

License Terms

SPT is noncommercial. You may not use this work for commercial purposes. For any reuse or distribution, you must make clear to others the license terms of this work. Any of these conditions can be waived if you get written permission from LVC/PUC-Rio.

If you use the SPT Tool in any of your experiments or research that leads to a scientific publication, please cite the following paper:

AYMA QUIRITA, VICTOR ANDRES; ACHANCCARAY DIAZ, PEDRO; FEITOSA, RAUL Q. ; HAPP, PATRICK N.; COSTA, GILSON A. O. P.; KLINGER, TOBIAS; HEIPKE, CHRISTIAN. Metaheuristics for Supervised Parameter Tuning of Multiresolution Segmentation. IEEE Geoscience and Remote Sensing Letters (Print), v. 13(9), p. 1364 – 1368, 2016.

ACHANCCARAY, P.; AYMA, V. A.; JIMENEZ, L. I.; BERNABE, S. G.; HAPP, P. N.; FEITOSA, R. Q. ; COSTA, G. A. O. P.; PLAZA, A. J. SPT 3.1: A free software tool for automatic tuning of segmentation parameters in optical, hyperspectral and SAR images. In: IEEE International Geoscience and Remote Sensing Symposium – IGARSS 2015, 2015, Milan. IEEE International Geoscience and Remote Sensing Symposium – IGARSS 2015, 2015. p. 4332-4335.

Achanccaray, P.; Ayma, V. A.; Jimenez, L. I.; Garcia, S. B.; Happ, P. N.; Feitosa, R. Q.; Plaza, A. SPT 3.0: A free software for automatic segmentation parameters tuning. In: Simposio Brasileiro De Sensoriamento Remoto, 17. (SBSR), 2015, João Pessoa, PB. Anais do 17mo Simpósio Brasileiro de Sensoriamento Remoto, João Pessoa, PB: INPE, 2015. p. 5578-5581.

Achanccaray, P.; Ayma, V. A.; Jimenez, L.; Garcia, S.; Happ, P.; Feitosa, R. Q.; Plaza, A.. A Free Software Tool for Automatic Tuning of Segmentation Parameter. South-Eastern European Journal of Earth Observation and Geomatics, V. 3(2S), P. 707-712, 2014.