[MISS 2016] Alison Noble - Learning to interpret Ultrasound Imaging скачать в хорошем качестве

[MISS 2016] Alison Noble - Learning to interpret Ultrasound Imaging

Lecture 2: Learning to interpret Ultrasound Imaging The acoustic patterns in ultrasound images contain rich information for clinical-decision making. However, from an image analysis perspective ultrasound is considered one of the hardest medical imaging modalities to work with due to the appearance being so different from MR/CT anatomical imaging and natural images, and the subjectivity of acquisition. Modern machine learning frameworks for interpreting images and video interpreting are well-suited to this imaging modality, and perhaps unsurprisingly their application is starting to transform thinking on how to automate analysis of ultrasound data. In this talk I will describe some of my groups inter-disciplinary research on machine learning applied to ultrasound. Challenges relate in part to how to automate detection and diagnostic measurement in data of variable quality typical of real world clinical use, but perhaps more scientifically interesting, how to design computer algorithms to recover biomarkers hidden within acoustic patterns that are difficult to perceive by the radiologists eye. Medical Imaging Summer School 2016 (31 July - 6 Aug 2016 Favignana, Sicily) - http://iplab.dmi.unict.it/miss16/