New Set of Algorithms to Create 3D Images of Cellular Structures using Light Microscopy

Our goal in this paper is to develop a practical, low-cost method for 3-dimensional volumetric imaging of cellular structures using a common light microscope. We have developed several algorithms that use 2-dimensional images obtained at different focus levels to construct a 3-dimensional image of the object by cleaning other layers' data from focused layer as much as possible. To extract clean raw images with eliminating point spread function's effects we used maximum likelihood and the Wavelet based hybrid regularization algorithm. We propose to map this 3D object extracted from deconvoluted images to a virtual space for high resolution viewing. This has the potential of greatly enhancing the viewing process and expose visual information which is otherwise hidden. Furthermore, with 3-D models available, the viewing process can be made more interactive allowing the scientists to study the specimens at different angles, cross-sections, zooming in/out, rotating the objects etc. The results illustrate that 3D object-from-focus imaging of translucent objects can be produced using light microscope and it is possible to effectively map that data to a virtual environment for enhanced user interactions.