2D-GE spot detection combining multidirectional texture and spatial intensity cues

Spot detection is a challenging task of 2D Gel Electrophoresis image analysis. The available software packages and techniques miss some of the protein spots while they detect a high number of spurious spots. This paper introduces a novel approach for the detection of protein spots on 2D gel images which is based on multidirectional texture and spatial intensity information. The proposed approach is compared with two commercial software packages using real 2D-GE images. The outcome demonstrates that the proposed approach outperforms the two software packages; it detects almost all of real protein spots and a low number of spurious spots.