Towards the automatic segmentation of HEp-2 cells in indirect immunofluorescence images using an efficient filtering based approach

The computer aided analysis of Indirect-Immunofluorescence (IIF) images is important for the differential diagnosis of several autoimmune diseases. A fully automatic approach consists in segmentation of individual cells in IIF images and subsequently its classification into various pattern types. This paper explores the segmentation of HEp2 cell in IIF images through the use of a filtering based approach. Our algorithm is based on a local convergence filter named as Sliding Band Filter (SBF). We propose a modified SBF that is capable of handling the low contrast, noise and illumination variations peculiar to IIF images. In addition, we follow a simple algorithmic pipeline and achieve better accuracy as compared to several state of the art segmentation algorithms on standard HEp2 image dataset.