Full text
| |
| Author(s): |
Celso Tetsuo Nagase Suzuki
Total Authors: 1
|
| Document type: | Master's Dissertation |
| Press: | Campinas, SP. |
| Institution: | Universidade Estadual de Campinas (UNICAMP). Instituto de Computação |
| Defense date: | 2007-02-27 |
| Examining board members: |
Alexandre Xavier Falcão;
Nelson Delfino d'Avila Mascarenhas;
Ricardo da Silva Torres
|
| Advisor: | Alexandre Xavier Falcão |
| Abstract | |
Segmentation is a process to partition an image into relevant regions for a given application. Segmentation is a fundamental stage of pattern recognition systems, being responsible for the extraction of objects whose features are used for recognition. In this work we developed a segmentation method, with high sensitivity and speci?city, to automatically separate human intestinal parasites (protozoa and helminth) from the background and impurities, like debris, food residue, animal and vegetal cells. The image collection used consists of intestinal parasites with high prevalence in Brazil. The proposed method reached a diagnostic sensitivity of 97,36% in detecting parasites, using a new pattern classi?er based on optimum-path forest. This work is a contribution to the development of an automatic diagnostic system for human intestinal parasites, to signi?cantly reduce the error rate in diagnostic by human inspection (AU) | |