Philip Batchelor

Philip Batchelor, (b. 30 December 1967 in St Austell, Cornwall, UK – d. 30 August 2011 near Annecy, France), was a Swiss-British academic in the fields of mathematics and medical imaging.

Life

Batchelor was born in in St Austell, Cornwall and grew up in Vouvry, a village in Switzerland. He graduated from ETH Zurich with a Masters in Theoretical Physics in 1992 and continued studying in Zurich for a PhD in Mathematics, which he obtained in 1997. In 1998, Batchelor joined United Medical and Dental School, which later became part of Kings College London (KCL), to work on the application of mathematical principles to magnetic resonance imaging (MRI) at Guy’s Hospital. He rapidly made an impact by employing concepts from differential geometry to study folding and curvature in the developing human brain.[1] In 2005 he moved to the Centre for Medical Image Computing at University College London, and was appointed a Senior Lecturer in the Imaging Sciences Division at KCL in 2006.

He gained a reputation for an ability to get to the heart of the issue and to draw in knowledge from other areas. A good example of this came during his work on diffusion tensor imaging (DTI). In trying to quantify the shapes of fibre tracts seen in DTI, he found measures that focussed on shape, not brain size.[2][3] Furthermore, these measures were inspired from such varied fields as torsion and polymer structure analysis and even the orbits of asteroids. In 2002, the choice of which diffusion gradient directions to acquire in diffusion tensor imaging was the subject of debate in the literature, with various empirical approaches having been proposed. At its core, this was a problem that combined the geometry related to the equal distribution of points on a hemisphere (the diffusion directions) and the diffusion-attenuated signal in MRI. Batchelor showed why direction schemes based on icosahedral shapes are optimal, and provided proof that the noise propagation in these icosahedral direction schemes was closely related to the best that could ever be done by acquiring an infinite number of directions.

There was also the issue of how to manipulate the tensor data from DTI. In one of his most highly cited papers,[4] Batchelor provided a rigorous framework that showed how to measure the geodesic distance between tensors (always maintaining an ellipsoid shape), how to average them, how to perform interpolation and rotation and he devised a metric called the geodesic anisotropy – an alternative to the commonly used fractional anisotropy measure.

Another key insight came from considering the key problem of how to correct non-rigid liver motion in MRI. Batchelor had the insight that this highly non-linear process could nonetheless be expressed concisely and in a general form using matrices.[5] This has opened the field of non-rigid motion correction in MRI (e.g.[6]).

Batchelor was a committed teacher and was always willing to patiently assist his colleagues and students with tutorials in mathematics. He recognised the need for education within the medical imaging academic community and in 2007 he organised a highly successful “Maths for Medical Imaging” summer school, which has since formed the basis of an ongoing component of the Kings College Medical Physics MSc.

His scope of research extended further including the visionary aim of producing diffusion tensor reconstructions of the beating heart. He and his group were the first to present 3D fibre images of the in-vivo heart of humans.[7] Compressed sensing also attracted his interest more recently. The work on k–t group sparse methods has significantly pushed the acceleration limits for dynamic MRI of the heart and lung.[8][9] He co-authored a number of other papers [10][11][12][13][14][15][16][17]

His characteristics of getting to the heart of the problem and bringing in fresh ideas were mirrored in his personality. Philip did not work for personal gain, he was open and selfless – committed to open source software and free access publishing. He was a skilled climber and loved mountain sports. He was killed in a climbing accident on 30 August 2011.

References

  1. ^ Batchelor PG, Castellano Smith AD, Hill DL, Hawkes DJ, Cox TC, Dean AF. Measures of folding applied to the development of the human fetal brain. IEEE Trans Med Imaging. 2002 Aug;21(8):953–65. PMID 12472268.
  2. ^ Batchelor PG, Atkinson D, Hill DL, Calamante F, Connelly A. Anisotropic noise propagation in diffusion tensor MRI sampling schemes. Magn Reson Med. 2003 Jun;49(6):1143–51. PMID 12768593.
  3. ^ Batchelor PG, Calamante F, Tournier JD, Atkinson D, Hill DL, Connelly A. Quantification of the shape of fiber tracts. Magn Reson Med. 2006 Apr;55(4):894–903. PMID 16526017.
  4. ^ Batchelor PG, Moakher M, Atkinson D, Calamante F, Connelly A. A rigorous framework for diffusion tensor calculus. Magn Reson Med. 2005 Jan;53(1):221–5. PMID 15690523.
  5. ^ Batchelor PG, Atkinson D, Irarrazaval P, Hill DL, Hajnal J, Larkman D. Matrix description of general motion correction applied to multishot images. Magn Reson Med. 2005 Nov;54(5):1273–80. PMID 16155887.
  6. ^ Atkinson D, Counsell S, Hajnal JV, Batchelor PG, Hill DL, Larkman DJ. Nonlinear phase correction of navigated multi-coil diffusion images. Magn Reson Med. 2006 Nov;56(5):1135–9. PMID 16986111.
  7. ^ Toussaint N, Sermesant M, Stoeck CT, Kozerke S, Batchelor PG. In vivo human 3D cardiac fibre architecture: reconstruction using curvilinear interpolation of diffusion tensor images. Med Image Comput Comput Assist Interv. 2010;13(Pt 1):418–25. PMID 20879258.
  8. ^ Usman M, Prieto C, Odille F, Atkinson D, Schaeffter T, Batchelor PG. A. Computationally efficient OMP-based compressed sensing reconstruction for dynamic. MRI. Phys Med Biol. 2011 Apr 7;56(7):N99–114. Epub 2011 Mar 2. PMID 21364267.
  9. ^ Usman M, Prieto C, Schaeffter T, Batchelor PG. k–t group sparse: a method for accelerating dynamic MRI. Magn Reson Med. 2011 Mar 9. doi:10.1002/mrm.22883 PMID 21394781.
  10. ^ Odille F, Uribe S, Batchelor PG, Prieto C, Schaeffter T, Atkinson D.Model-based reconstruction for cardiac cine MRI without ECG or breath holding.Magn Reson Med. 2010 May;63(5):1247–57. PMID 20432296.
  11. ^ Boubertakh R, Prieto C, Batchelor PG, Uribe S, Atkinson D, Eggers H, S√∏rensen TS, Hansen MS, Razavi RS, Schaeffter T. Whole-heart imaging using undersampled radial phase encoding (RPE) and iterative sensitivity encoding (SENSE) reconstruction. Magn Reson Med. 2009 Nov;62(5):1331–7. PMID 19780159.
  12. ^ Knowles BR, Batchelor PG, Parish V, Ginks M, Plein S, Razavi R, Schaeffter T. Pharmacokinetic modeling of delayed gadolinium enhancement in the myocardium. Magn Reson Med. 2008 Dec;60(6):1524–30. PMID 19025896.
  13. ^ Prieto C, Batchelor PG, Hill DL, Hajnal JV, Guarini M, Irarrazaval P. Reconstruction of undersampled dynamic images by modeling the motion of object elements. Magn Reson Med. 2007 May;57(5):939–49. PMID 17457881.
  14. ^ Larkman DJ, Batchelor PG, Atkinson D, Rueckert D, Hajnal JV. Beyond the g-factor limit in sensitivity encoding using joint histogram entropy. Magn Reson Med. 2006 Jan;55(1):153–60. PMID 16342149.
  15. ^ Penney GP, Batchelor PG, Hill DL, Hawkes DJ, Weese J. Validation of a two- to three-dimensional registration algorithm for aligning preoperative CT images and intraoperative fluoroscopy images. Med Phys. 2001 Jun;28(6):1024–32. PMID 11439472.
  16. ^ Atkinson D, Larkman DJ, Batchelor PG, Hill DL, Hajnal JV. Coil-based artefact reduction. Magn Reson Med. 2004 Oct;52(4):825–30. PMID 15389945.
  17. ^ Hill DL, Batchelor PG, Holden M, Hawkes DJ. Medical image registration. Phys Med Biol. 2001 Mar;46(3):R1–45. Review. PMID 11277237.