Outline Detection for the HEp-2 Cell in Indirect Immunofluorescence Images Using Watershed Segmentation

An automatic inspection system for antinuclear autoantibodies (ANA) testing can be divided into HEp-2 cell detection, fluorescence pattern classification and computer aided diagnosis phases. This paper presented a multi-staged segmentation method for automatically detecting outlines of fluorescence cells in Indirect immunofluorescence (IIF) images. The similarity-based watershed algorithm performed the marker to prevent over-segmentation. This paper evaluated 2305 autoantibody fluorescence patterns with manual sketched outlines (including 456 diffuse patterns, 417 peripheral patterns, 719 coarse speckled patterns, 55 fine speckled patterns, 517 discrete speckled patterns and 141 nucleolar patterns) from 44 IIF images. The experimental results revealed that the proposed method can practically outline fluorescence cells from IIF images.