Improved signal extraction from fluorescence immunoassay image sequences

Signal extraction from fluorescence immunoassay image sequence captured by charge coupled device arrays suffers from object-dependent inhomogeneous spatial gain. In a recently studied fluorescence immunoassay for myoglobin, a potential biomarker for heart attack, this problem results in small signal-to-noise ratio of the extracted signal which in turn degrades the detection limits and estimation consistency.; The problem is analyzed from an image processing perspective whereby the image distortion is studied and modeled. A novel correction method is then proposed which restores an image by estimating the distortion from the image itself. Signal extraction by segmentation from the restored image improves the fluorescence signal content. Enhanced detection limits and improved trends in the consistency of estimated concentrations of the biochemical are also observed showing the possibility of an accurate and time-efficient method of biochemical concentration estimation in sub-nanomolar levels.