Software development in medical image processing - Good practices and methodology

Medical imaging covers the acquisition and reproduction of images of the human body based on various physical phenomena such as X-ray absorption, nuclear magnetic resonance, ultrasound wave reflection, radioactivity or, more simply, optical imaging.

It has revolutionized medicine by enabling non-invasive, dynamic visualization of the anatomy, physiology and metabolism of the human body for diagnostic purposes. This indirect visualization requires the use of image processing techniques to detect, quantify, characterize or filter the digital data acquired.

Successfully implementing these techniques is a highly collaborative and often translational engineering task, all within the critical context of preserving patient safety.

Over and above image processing techniques and methods, this article will attempt to provide some methodological elements to help such work achieve its objective, in an industrial or research context.

Our aim is to provide engineers, researchers and technical supervisors with concrete information, such as steps, tools and simple questions to ask. Bibliographical references will enable you to delve deeper into more fundamental or general aspects. Finally, the notions of context and purpose will be analyzed in particular, to give the project manager a sense of perspective and enable him or her to adjust efforts to the intended target.