• digital mammography
• GaAs detectors
• image quality
• Single Photon Counting

The work presented in this thesis is about an innovative digital mammographic system developed in the framework of the research project “Integrated Mammographic Imaging” (IMI). This prototypal device is the first full-field mammographic imaging system based on GaAs pixel detectors and on Single Photon Counting (SPC) read-out electronics.
The GaAs sensors developed for the IMI prototype are composed of a 200μm thick Semi-Insulating GaAs crystal on which a matrix of 64 × 64 square electrodes (pixels), 150μm in side and with 20μm of inter-pixel distance, is deposited. The electrodes are connected by bump-bonding to an equally segmented read-out chip, named Medipix1. The Medipix1 chip works in SPC mode, for which the noise can be totally rejected by setting a discrimination threshold above the noise level and by counting only the photons with energy over threshold.
Each assembly, constituted by the GaAs sensor coupled to Medipix1, is a basic detection unit of about 1.2cm2 of active area. The IMI prototype uses 18 of these detection units mounted onto a chip-board and scanned to cover the standard (18 × 24)cm2 exposure area.
The aim of this thesis is to perform a complete characterization of the IMI protptype and to assess its image quality capabilities.
In the first part of this work is described the calibration and optimization of the system for mammographic imaging use.
Furthermore we have evaluated the spectral capabilities of the detectors by means of Monte Carlo simulations and experimental spectroscopic measurements.
In order to assess the imaging capabilities of the IMI system, two methodologies have been followed. The first one is based on the experimental measurement of the transfer functions of the system. The Modulation Transfer
Function (MTF), the Noise Power Spectrum (NPS) and the Detective Quantum Efficiency (DQE) are figures of merit that describe the spatial resolution, noise and SNR transfer capability of the system.
This analysis allows evaluating in an objective way the performance of the X-ray detector, but it does not take into account other aspects of the image visualization. For this reason another approach, based on the Contrast Threshold (CT) analysis has been also used. This analysis takes into account also the visual processes of the human observer.