Three-dimensional deconvolution of optical microscope images using adaptive inverse modelling

Three dimensional microscope imagery has unavoidable blurring due to out-of-focus and convolutional effects that inhibit resolution of submicron detail in biological cells. In this study, a three dimensional FIR filter is used to adaptively inverse model the point spread function that characterizes the aforementioned corruption. A submicron sized microbead, essentially an optical impulse function, is imaged and used in an adaptive inverse modelling configuration with the desired signal taken as an approximation to an impulse (an exponential that decreases rapidly). The resulting inverse filter is then applied to the data from the microscope. The results of this overall scheme are superior to the presently used spatial frequency techniques that introduce distortion due to frequency domain artifacts.