Evaluation of the quality of ultrasound image compression for a robotic tele-echographic system

The subject of this paper is to propose a method for the evaluation of ultrasound image compression. Compression techniques are used for image transmission in the frame of two tele-operated robotic chains dedicated to tele-echography: OTELO and TERESA. Our objective is to define a statistical criterion to measure the image quality with the same reliability than the one provided by a medical assessment. An initial psychovisual experiment is proposed to medical experts, and represents our reference value for the comparison of statistical evaluation criteria. We propose to fusion different some statistical criteria and to exploit the medical expert judgments during a training phase. Two methods are tested for this approach: linear combination of the criteria with a genetic algorithm and learning the experts judgment by a support vector machine. We show the benefit of this methodology through some experimental results