Linear Local Distance coding for classification of HEp-2 staining patterns

Abstract

Indirect Immunofluorescence (IIF) on Human Epithelial-2 (HEp-2) cells is the recommended methodology for detecting some specific autoimmune diseases by searching for antinuclear antibodies (ANAs) within a patient's serum. Due to the limitations of IIF such as subjective evaluation, automated Computer-Aided Diagnosis (CAD) system is required for diagnostic purposes. In particular, staining patterns classification of HEp-2 cells is a challenging task. In this paper, we adopt a feature extraction-coding-pooling framework which has shown impressive performance in image classification tasks, because it can obtain discriminative and effective image representation. However, the information loss is inevitable in the coding process. Therefore, we propose a Linear Local Distance (LLD) coding method to capture more discriminative information. LLD transforms original local feature to local distance vector by searching for local nearest few neighbors of local feature in the class-specific manifolds. The obtained local distance vector is further encoded and pooled together to get salient image representation. We demonstrate the effectiveness of LLD method on a public HEp-2 cells dataset containing six major staining patterns. Experimental results show that our approach has a superior performance to the state-of-the-art coding methods for staining patterns classification of HEp-2 cells.