Computational Optics and Vision Engineering
The Computational Optics and Vision Engineering Research Theme brings together academics within different Groups in the Department who share an interest in imaging technology, microscopy, image and video processing, computer vision, information processing/analysis (including high-level image analysis and understanding, and machine learning) and inverse problems related to image and diffraction data.
The applications of our work are diverse:
- ultra-high resolution electron microscopy and electron diffractive imaging
- novel methods for 3D semiconductor lithography
- x-ray and optical lensless imaging (ptychography)
- biomedical imaging
- image data compression and image restoration/enhancement
- object and scene recognition, human activity analysis, industrial inspection problems
- time-domain signals.
The computational methods we employ – data compression, compressive sensing, wavelet transforms, dictionary learning, pattern classification, high-level image processing, Fourier methods, phase-retrieval, advanced optimisation and inverse methods – are common to many of our interests.
Research areas and people
- Optical, x-ray and electron lensless imaging: ptychography
- Computer Vision
- Machine learning
- Ultra-high resolution electron microscopy
- 3D lithography
PhD Opportunities available:
What is lensless imaging?
When you think of a microscope, you usually think of looking through lenses. The lenses magnify the object and focus on details within it.
We make transmission (and reflection) microscopes without using lenses. Instead, we record the intensity of the radiation scattered by the object (the diffraction pattern) and process this computationally to form the image. Unfortunately, the phase of the scattered wave (the time of arrival of the peaks and troughs of the waves) is lost when we measure intensity. To get this back (essential if we need to form an image), we need to undertake an elaborate inverse calculation.
We use this roundabout method because for short wavelength radiation x-rays and electrons) it’s very hard to make a good quality lens. Lensless imaging disposes of all the problems associated with lenses. As it happens, a lensless image has more information in it than a conventional image taken with a good quality lens! So lensless imaging is useful even at optical wavelengths.
Ptychography
The particular method we work on is called ‘ptychography’. It processes many diffraction patterns taken from different areas of the object. The technique originally developed by Prof John Rodenburg in the Electronics and Electrical Engineering Department in the University of Sheffield is by far the most effective lensless imaging approach. It has been adopted widely in the x-ray imaging community and is now available commercially at optical wavelengths. We are now actively developing the method at electron wavelengths. Prof John Rodenburg FRS and Dr. Andrew Maiden give a brief overview of how Ptycography works in the videos below. For more information please contact: Professor John Rodenburg, or Dr Andrew Maiden.
Humphrey, M.J., Kraus, B., Hurst, A.C., Maiden, A.M., and Rodenburg, J.M.
Ptychographic electron microscopy using high-angle dark-field scattering for sub-nanometre resolution imaging'
Nature Communications 3, Art. No: 730 doi:10.1038/ncomms1733
For more information, contact: