Microarray Digital Image And Information Processing

Microarray images are acquired by scanning the fluorescence excited from the genechips after hybridization. Through spotting, extracting the intensities for every probe, then calculating mean intensities and ratios, digital image processing of microarray provides biologists with essential conditions for further research.This paper focuses on three parts, filtering image, spotting and gridding, finally normalization. Corresponding to each part an improved method was proposed.A new method based on mathematical morphology for the gridding of microarray images is put forward, which is characterized by filtering out the block noise, filtering projection plots and ending up with automatic correction. It has been proved that images heavily contaminated by noises tend to be processed correctly, which greatly increases the success rate of gridding.In this paper an improved normalization method, Q-LOWESS...