Imaging with Rays: Microscopy, Medical Imaging, and Computer Vision
This work addresses imaging challenges in microscopy, medical imaging, and computer vision, but appears incremental as it extends existing tomographic reconstruction methods to handle occlusion.
The paper tackles the problem of imaging opaque objects with occlusion by formulating reconstruction as a regularized nonlinear optimization to solve for brightness and attenuation, demonstrating simulated examples including a super-resolving technique that exploits occlusion.
In this paper we broadly consider techniques which utilize projections on rays for data collection, with particular emphasis on optical techniques. We formulate a variety of imaging techniques as either special cases or extensions of tomographic reconstruction. We then consider how the techniques must be extended to describe objects containing occlusion, as with a self-occluding opaque object. We formulate the reconstruction problem as a regularized nonlinear optimization problem to simultaneously solve for object brightness and attenuation, where the attenuation can become infinite. We demonstrate various simulated examples for imaging opaque objects, including sparse point sources, a conventional multiview reconstruction technique, and a super-resolving technique which exploits occlusion to resolve an image.