Francesco Verdoja

Role: PhD Student
Address: Ufficio 14 ex lab
Dipartimento di Informatica - Università degli Studi di Torino
Via Pessinetto 12, 10149 Torino, Italy icona mappa
E-mail:
Telephone: n.d.
Skype: frasmog
Reception: on appointment

foto profilo docente

icona biografiaShort Bio

Francesco Verdoja got his master degree in Computer Science with full marks at University of Turin in October 2013. During his master thesis he contributed to a biomedical research project involving IRCC (Institute for Cancer Research and Treatment) and Computer Science Dept. The project was devoted to the design of novel and more accurate automatic tumor detection and segmentation methods in PET images. In this framework, he has proposed a novel automatic method for tumor segmentation in dynamic PET images. The method is based on registration of multiple 3D dynamic PET scans and innovative application of the RX detector to isolate the tumor voxels. The proposed method exhibits promising experimental results and has been presented in the IEEE Int. Conference of Image Processing (ICIP 2014). Moreover, the method has raised the interest of the medical personnel of the IRCC for its potential applications to radiotherapy.

At present, Francesco Verdoja is pursuing his Ph.D degree in Computer Science at University of Turin under the supervision of Prof. Marco Grangetto.

His Ph.D program is part of a funded research project devoted to the investigation of image segmentation and advanced computer vision algorithms in light of their application to future image and video compression techniques. The project involves as partners University of Turin, Polytechnic University of Turin, Italy's national public broadcasting company (RAI) and Sisvel Technology. The research is expected to contribute to future International standards.

Currently, he is doing research on superpixels and clustering techniques based on graphs of superpixels. The goal is the design of a novel image segmentation method jointly taking into account segmentation accuracy and rate/distortion costs; a graph based approach has been chosen since it is likely to serve as unifying tool for segmentation, description and compact representation of visual data.

Download CV in PDF (ENG) icona file pdf
Scarica CV in PDF (ITA) icona file pdf