Noncontact optical tomography in reflection mode is often the only possible configuration when imaging the expression of green fluorescent protein (GFP) or other fluorescent proteins in live animals owing to the short penetration depth of visible light. When imaging in reflection mode using noncontact approaches (i.e., without the use of fibers coupled to tissue), correctly accounting for the intensity profile of the source at the surface is a difficult task, usually needing to fit for source positions and/or approximating these to point sources. In this Letter we present a rigorous theoretical approach that directly accounts for the source's intensity profile and verify it using in vivo data from GFP-expressing mice. We show how this approach improves image quality and resolution, while considerably simplifying the forward and inverse problems of the image reconstruction process.
© 2009 Optical Society of AmericaPDF Article