Analysis Of Protein Spots Detection Approach Based On Fuzzy Morphological And Watershed

The extraction and analysis of difference protein point in two-dimensional gelimage is the core content of gel electrophoresis technology, and currently gelelectrophoresis technology is one of the important methods in proteomics research, thistechnology can not only separate protein point effectively, but also be used for thequalitative and quantitative researching of protein point. This paper combined with thenational natural science funds projects and jiang xi province natural science fundprojects, took the protein spots detection in two-dimensional gel image as theimportance of research, the main research results and contributions are concluded asfollows:(1) Had a systematically studying in the relevant theoretical knowledge of proteinspots detection of gel image. This article expounded all kinds of image preprocessingmethods and principles, traditional watershed transformation algorithm principle andthe problems of the improved watershed transform algorithm, proposed a markercontrol watershed algorithm, which covers several advantages in other watershedalgorithms mentioned before is simple and effective.(2) For the problem of over segmentation, edge discontinuous and missing in theprotein spot detecting, a protein spot area detection approach based on maker controlwatershed was presented. Firstly, the common and improved filter method were used toavoid the interference of stripe, crack and background transform; Secondly makercontrol watershed was applied twice to reduce the fake extreme value point, most ofnon protein spot area was removed effectively; Finally, in order to get qualified proteinspot area, this paper put forward a criterion for judging the qualified protein spot area.Many gel image experimental results showed that this approach could detect the mostprotein spot area, avoid the problem of the image segmentation, edge discontinuousand protein point area missing, which has good robustness and practical.(3) For the problem of overlapping in the protein spot detecting, a overlappedprotein spot segmentation method based on maker control watershed was proposed.connected component labeling and distance transformation were supplied to extract theoverlapped protein spot area, and watershed was used for the third time to separate theoverlapped protein spot accurately. Experimental results on overlapped protein spot area gel image showed that this method which could solve the overlapping problem,had high efficiency and practicability.