FHIS Image Segmentation Library

Andrea Pennisi

Fast Hsv Image Segmentation (FHIS) Library is an OpenCV based C++ adaptation of the original Matlab code designed for performing an accurate segmentation in real-time. FHIS creates a simple representation of the image by using the Delaunay triangulation. A HSV threshold is used in order to find similar triangles, obtaining an accurate segmented image. FHIS exploits OpenCV 2 and CGAL functions. The library is based on the work [1] realized by Camillo Taylor and Anthony Cowley of University of Pennsylvania.





A detailed description of the original segmentation method by Taylor and Cowley can be found in the paper "Parsing Indoor Scenes Using RGB-D Imagery" (pdf) in the proceedings of Robotics: Science and Systems (Science and System) 2012.

Please, cite the above paper and this page if you use FHIS.




FHIS is provided without any warranty about its usability. It is for educational purposes and should be regarded as such.



FHIS Prerequisites



Build FHIS

FHIS is provided with a CMakeLists.txt file and can be compiled by using CMake.


  1. Unzip the file fhis.zip in
  2. $cd
  3. $mkdir build
  4. $cd build
  5. $cmake ..
  6. $make



FHIS is provided with an usage example (main.cpp)


For video files

$./fhis -vid video1.avi

For a single image:

$./fhis -singleimg images/1.png

For an image sequence (fps = 25 default value)

$./fhis -img images/1.png

In addition you can specify the fps value

$./fhis -img images/1.png -fps 7

Note that the usage example works only on image sequences in which the filename format is .png, where n is the frame number (e.g., 7.png).




FHIS has been tested with multiple well-known image sequences representing different scenarios.

Results obtained by using FHIS are reported in the following.

  1. Indoor Environments [1]    raw data    fhis results
  2. PETS [2]    raw data    fhis results




FHIS has been written by Andrea Pennisi.

The author thanks Dr. Domenico Daniele Bloisi.




[1] Camillo Taylor and Anthony Cowley, "Parsing Indoor Scenes Using RGB-D Imagery", Proceedings of Robotics: Science and Systems, July 2012.

[2] PETS Data set: http://www.cvg.rdg.ac.uk/PETS2009/.