A level set approach for proteomics image analysis

This paper addresses the detection of protein spots in pro-teomics images. A novel level set approach is proposed, which is capable of separating spots belonging to multi-plets. The proposed approach copes with the presence of noise and the inhomogeneous background, which are major issues in proteomics image analysis. It has been experimentally compared with a recently published method for spot detection. The results of the experiments performed indicate that the previously proposed method tends to falsely identify multiplets as singlets, whereas the proposed approach succeeds in spot separation.