The Cambridge software package for image reconstruction from optical interferometry data has been dubbed BSMEM, to stand for BiSpectrum Maximum Entropy Method. It was first written in Fortran and completed by David Buscher in 1992 to demonstrate direct maximum entropy reconstruction from optical aperture synthesis data (Buscher 1994). BSMEM applies a fully Bayesian approach to the inverse problem of finding the most probable image given the evidence. BSMEM makes use of the MemSys library created by Maximum Entropy Data Consultants Ltd. to implement a gradient descent algorithm for maximising the inference (posterior probability) of an image, using the entropy of the reconstructed image as the prior probability.
From 2002 to 2006, while the core of the program has remained more or less unchanged, multiple improvements have been implemented by Hrobjartur Thorsteinsson and Fabien Baron. BSMEM has partly been ported to C, it has gained compatibility with the OIFITS format, and the reconstruction options and the user interface have been improved and expanded.
We acknowledge funding from the European Community’s Sixth Framework Programme under contract number RII3-CT-001566 (OPTICON JRA-4), which has contributed to recent development of BSMEM.
Features of BSMEM:
- Fully Bayesian approach
- Handles all types of data sparseness (missing power spectra and/or missing bispectrum amplitudes and/or missing bispectrum phases)
- Speed: reconstruction of a 256 x 256 pixel image from a typical medium signal-to-noise interferometric dataset takes only a few minutes on a modern computer
- Reads the OIFITS data exchange format
Papers and Reports on BSMEM
- Direct maximum-entropy image reconstruction from the bispectrum, D. F. Buscher. In J. G. Robertson and W. J. Tango, editors, IAU Symp. 158: Very High Angular Resolution Imaging, p.91 (1994).
- An interferometry imaging beauty contest, P. R. Lawson, W. D. Cotton, C. A. Hummel, J. D. Monnier, M. Zhao, J. S. Young, H. Thorsteinsson, S. C. Meimon, L. Mugnier, Guy Le Besnerais, E. Thiébaut, P. G. Tuthill, Proc. SPIE 5491, New Frontiers in Stellar Interferometry, pp.886-899 (2004) – PDF.
- 2006 interferometry imaging beauty contest, P. R. Lawson, W. D. Cotton, C. A. Hummel, F. Baron, J. S. Young, S. Kraus, K.-H. Hofmann, G. Weigelt, M. Ireland, J. D. Monnier, E. Thiébaut, S. Rengaswamy, O. Chesneau, Proc. SPIE 6268, Advances in Stellar Interferometry (2006) – PDF.
- BSMEM Test Report, F. Baron, EII-JRA4 document JRA4-TRE-2520-0002 (2007).
- 2008 imaging beauty contest, W. Cotton, J. Monnier, F. Baron, K.-H. Hofmann, S. Kraus, G. Weigelt, S. Rengaswamy, E. Thiébaut, P. Lawson, W. Jaffe, C. Hummel, T. Pauls, H. Schmitt, P. Tuthill, J. Young, Proc. SPIE 7013, Optical and Infrared Interferometry (2008) – PDF.
- The 2010 interferometric imaging beauty contest, F. Malbet, W. Cotton, G. Duvert, P. Lawson, A. Chiavassa, J. Young, F. Baron, D. Buscher, S. Rengaswamy, B. Kloppenborg, M. Vannier, L. Mugnier, Proc. SPIE 7734, Optical and Infrared Interferometry II (2010) – PDF.
BSMEM contains proprietary software, but is available for academic use subject to a no-fee licence agreement.
There is currently very little documentation distributed with BSMEM. Hence we recommend that only those with extensive experience of interferometric image reconstruction attempt to use BSMEM at the moment.
To obtain BSMEM, please download the licence agreement linked below, print it out, sign it, and mail the signed copy to:
Maximum Entropy Data Consultants Ltd.
South Hill, 42 Southgate Street
Bury St. Edmunds
Suffolk IP33 2AZ
Please email John Young (see our staff page for address) once you have sent off the licence agreement; this will expedite the process.
BSMEM Academic Licence Agreement (PDF format)