For the licentiate degree in Electrical Engineering thesis, I created an image segmentation algorithm called PDF Image Segmentator. This algorithm is capable of segmenting an object from its background in an intensity image (given that the object is distinguishable by its gray levels from the background), using a maximum likelihood thresholding method based on four-parameter gamma distributions. The algorithm works very good even for low contrast images, and can be used for segmenting objects in many applications: like cell segmentation and handwriting recognition.

 

My tutor, Prof. Geovanni Martinez, PhD., guided me in a great way.

 

A scientific article written was accepted in the 2014 11th International Conference on Electrical Engineering, Computing Science and Automatic Control 2014 (CCE 2014, Ciudad del Carmen, Mexico, October 2014), titled: “Maximum Likelihood Thresholding Algorithm Based on Four-Parameter Gamma Distributions”.

 

I presented a poster about this research in the Hands-On Research in Complex Systems School at the International Centre for Theoretical Physics (ICTP), Trieste, Italy, 1 – 12 July 2013, and won a Jury’s Choice Poster Award.

 

(Thesis document)(Thesis presentation)(Poster)

 

Image Processing and Computer Vision Research Laboratory (IPCV-Lab) of the University of Costa Rica

February 2011 - February 2014

 

Reference:

Geovanni Martinez

gmartin@eie.ucr.ac.cr

Image Segmentator