Centre for Medical Image Computing (CMIC)


Patrice Lambert - Colin Reveley

Start: Feb 28, 2018 01:00 PM
End: Feb 28, 2018 02:00 PM

Location: UCL Bloomsbury - Roberts 106 Roberts building

Patrice Lambert

Title: Parallel Robots with Configurable Platforms


Parallel robots are formed by two rigid links, the base and the platform, connected in parallel by independent serial chains, called legs. Usually, one joint per leg near the base is actuated to provide full control of the platform. Since they offer high structural stiffness, they are widely used in applications that required high mechanical bandwidth, such as flight simulators, pick-and-place manipulators, and haptic interfaces.

The new concept behind parallel robots with configurable platform is that the rigid platform is replaced by a closed-loop chain, a configurable platform. Some of the mechanical links of this closed-loop chain are attached to the legs, so both the motion and the configuration of the platform can be fully controlled by the actuators of the legs. The use of a configurable platform allows the robot to interact with the environment via multiple contact points. This results in a robot that can combine motion and grasping capabilities into a structure that provides an inherent high structural stiffness.

General notions about parallel robots with configurable platforms will be introduced, and the design and implementation of some prototypes based on this concept will be presented.

Colin Reveley, Department of informatics, Sussex University

Title: dMRI Correlates of Gray Matter Structure : Pilot Work


Historically, it has been thought that diffusion in gray matter is isotropic. More recently, evidence of changes in anisotropy in the cortex from the white matter boundary to the pial surface has accumulated in-vivo. This talk shows pilot work with ultra-high resolution ex-vivo dMRI, in which we demonstrate that diffusion tensors in gray matter show rich variety. We link this variety to subtle anatomical features using histology from the same brains. It seems likely that diffusion imaging is capable of identifying anatomical regions (like motor or sensory) in the cortex, or detecting markers for changes in tissue structure caused by diseases in a way that pushes beyond measures of cortical myelin content. At present my colleagues and I are about to start data acquisition for a major project in the NIH. We will be scanning ex-vivo marmoset brain samples at very high field for many days at a time, and then sectioning them for microscopy. We are very interested in hearing new ideas for MRI sequences, models, tissue stains and especially image processing methods that might form a quantitative bridge between our histology data and our MRI.

Short bio:

I'm a computational neuroanatomist. I am associate faculty in the department of informatics at Sussex University and guest researcher at the Neurophysiology Imaging core facility at the National Institutes of Health in Bethesda, Maryland USA.