2018 John William Strutt, Lord Rayleigh Medal and Prize

Dr Owen Saxton of Murray Edwards College, University of Cambridge for his contributions to the Gerchberg-Saxton computer algorithm, decades ahead of its time but now prevalent in phase retrieval, and for his foundational image processing programs, still influential in front line electron microscopy.

2018 John William Strutt, Lord Rayleigh Medal and Prize

Dr Owen Saxton devised many seminal advances in computer image processing that continue to underpin front-line electron microscopy. In his 2017 Nobel Lecture, Joachim Frank specifically recalled his early work (Saxton and Frank 1977) on the development of cross-correlation methods for motif detection in noisy images. For cryo-microscopy this was a crucial early step in the development of processing methods for identical objects, randomly arranged in different positions, orientations and strain configurations.

Saxton's work in 1982 with Wolfgang Baumeister, addressing the problem of strain in more periodic structures, had a more immediate impact through the correlation averaging of a bacterial cell envelope protein. This paper received more than 600 citations because it introduced the now commonly used Fourier Shell Correlation resolution criterion. In the 1986 development of their ‘lattice unbending’ technique for imaging the purple membrane, Henderson et al acknowledged their indebtedness to Saxton.

Through Synoptics, the company that Saxton and two colleagues founded, SEMPER – a comprehensive and powerful package of image processing programs – was made available for more general use in electron microscopy. SEMPER was very widely used until the 1990s, when the successor company discontinued its support, and still has its devotees.

With their computer algorithm for phase retrieval by cycling between images and diffraction patterns, Ralph Gerchberg (sadly deceased) and Saxton lit a truly remarkable slow-burning fuse. In successive decades their 1972 publication received 60, 200, 500 and 900 citations and now collects about 200 each year. Initial applications, mostly in electron microscopy and often limited to one dimension, failed to reveal the true capabilities of the algorithm. However, interest grew following James Fienup’s influential paper in 1982 assessing its wide applicability, stability and convergence and his 1993 use of it to correct the aberrations of the Hubble Telescope. The algorithm is now the key feature of most phase retrieval methods, including ptychography, and is being energetically applied in optical and X-ray imaging.

Saxton is a prime example of a brilliant thinker whose best work was far ahead of its time. Recognition of his achievements has been lamentably slow and continues to be inadequate.



Cookie Settings