Sorting and obtaining target biological species such as live cells from complex mixtures are critical for biotechnology research, diagnostics, and therapeutics. In laboratory, cell sorting is performed by fluorescence-activated cell sorting(FACS). However, FACS is inherently limited in sorting cells based on the observation of cellular features. Applications such as cell-based therapeutics, stem cell culturing, and in-cell directed evolution of functional macro-molecules all require the enrichment of target cells from complex mixtures. For these cases, image-based discrimination and sorting must be performed simultaneously at the single cell level. We propose to meet this need by utilizing recent advances in imaging technology and computer-based image processing.