A Novel Image Methodology for Interpretation of Gold Immunochromatographic Strip

Abstract

Gold immunochromatographic strip assay is a rapid, simple, single-copy and on-site method. Quantitative Interpretation of the strip can provide more information than the traditional qualitative or semiquantitative strip assay. The paper aims to develop an image based assay method for quantitative determination of trace concentrations by gold immunochromatographic strip. The image of gold immunochromatographic strip is taken by CCD, and, after the proper filter and window cutting, the test line and control line is segmented by the genetic fast fuzzy c-means(FCM) clustering algorithm. In order to improve the measure property, based on Lambert-beer law, the relative reflective integral optical density(RIOD) is selected as the feature by which the interference in the test and control lines can be canceled out each other. The proposed method is applied to the quantitative detection of human chorionic gonadotropin (hCG) as a model. Firstly, the segmentation performance of the genetic fast FCM clustering algorithm is compared with threshold method and FCM clustering algorithm in terms of the peak signal-to-noise ratio (PSNR). Furthermore, the comparison of the blind experiment between the proposed method and commercial quantitative instrument swp-sc1 is carried out. This method is shown to deliver a result comparable and even superior to existing techniques.