Microarray Images Processing Based on Mathematical Morphology

Noises in microarray images can be divided into pulse noise and block noise. In most researches block noise is tended to be left untreated while pulse noise has been eliminated, which leads to incorrect results in gridding. In this paper, an improved method based on mathematical morphology for the gridding of microarray images is proposed, which is characterized by filtering out the block noise and filtering projection plots. Processing results have shown that images heavily contaminated by noises tend to be processed correctly, which greatly increases the success rate of gridding