MHEDHBI Imen

PhD student at Sorbonne University
Team : SYEL
https://lip6.fr/Imen.Mhedhbi

Supervision : Patrick GARDA

Co-supervision : HACHICHA Khalil

Compression en qualité diagnostic de séquences d'images médicales pour des plateformes embarquées

Hospitals and medical centers produce an enormous amount of digital medical images every day, especially in the form of image sequences. Due to the large storage size and limited transmission bandwidth, an efficient compression technique is necessary. We first proposed a compression algorithm for medical images sequences MMWaaves. It is based on Markov fields coupled with the certified medical device Waaves of Cira company. We demonstrated that MMWaaves provide a compression gains greater than 30% compared to JPEG2000 and Waaves while ensuring outstanding image quality for medical diagnosis (SSIM> 0.98). In addition, it achieved compression rates equal to those obtained by H.264 while improving the image quality. Then we developed a new compression algorithm MLPWaaves based on DWT difference followed by a new adaptive scanning model LPEAM in order to optimize the local stationary of wavelet coefficients. We obtained a compression gain up to 80% compared to Waaves and JPEG2000 while ensuring acceptable quality for medical diagnosis. Finally, in order to transmit medical images for diagnostic from the health center to the mobile device of the doctor, we proposed a client-server remote radiology system for encoding and decoding. It is based on a multithreading paradigm to accelerate treatment. The validation of this solution was performed on two different platforms. We achieved an acceleration factor of 5 on an Intel Core i7-2600 and a factor of 3 on a Samsung Galaxy tablet.

Defence : 01/15/2015

Jury members :

Azedine Beghdadi (Professor L2TI - Institut Galilée Université Paris 13 – Sorbonne Paris Cité) [Rapporteur]
Rabah Hassan (Professeur, Université de Lorraine) [Rapporteur]
Aziz Bensrhair (Professeur, INSA de Rouen)
Matthieu Cord (Professeur, LIP6, Université Pierre et Marie Curie))
Sylvain Hochberg (Gérant société CIRA)
Didier Heudes (Docteur, HEGP)
Khalil Hachicha (Maitre de Conférence, LIP6, Université Pierre et Marie Curie)
Patrick Garda (Professeur, LIP6, Université Pierre et Marie Curie)

Departure date : 01/16/2015

2012-2017 Publications