David Atkinson is a Reader in Magnetic Resonance Imaging (MRI) at University College London. He has a degree in Physics and a PhD in the field of semiconductors and optoelectronics. After a period as a scientific programmer in industry and a PostDoc at UCL, he moved to Guy's Hospital, London in 1996 to work on motion correction in MRI. In 2005 he returned to UCL as part of the newly formed Centre for Medical Image Computing within the UCL Departments of Medical Physics and Computer Science. Since 2011 he has been based in the UCL Division of Medicine within the Centre for Medical Imaging and also holds an Honorary NHS position at University College London Hospital. He facilitates clinical imaging, clinical research studies and has his own areas of active research.
Magnetic Resonance Imaging (MRI) can provide excellent visualisation of different types of soft tissue and uses no ionising radiation. Within MR scanners, pulse programming software controls the timing of RF pulses and magnetic field gradients used to generate MR signals. These signals are received over time and then reconstructed to form an image. Dr Atkinson's research has investigated new sequences and post-processing algorithms to infer clinically-relevant information about tissue properties. Research has included characterisation of tumours, placental and brain tissue, the removal of image artefacts, reduced distortions in prostate cancer imaging, and, the use of motion to characterise diseases of the gastro-intestinal tract. Motion at any time during an MR acquisition can result in blurring and ghosting across the reconstructed images. Convolutional Neural Networks can learn image features and the Turing Fellowship will be used to investigate the application of these networks for the removal of motion artefacts in MR images.