| 摘要 |
A microarray image is segmented into discreet segments for respective spots in the microarray image based on locating microarray image rows and columns by maximizing scores associated with vertical and horizontal image projections. To determine the locations of the rows and columns that extend through artifacts that involve multiple spots and/or multiple columns and rows, the scores may also include contributions from the gaps between adjacent rows and columns. The scores are calculated by stepping windows containing nominally spaced spot-sized row or column boundaries, over the horizontal projection curves. At each step a total row or column score is calculated that represents the areas of the projection curves that fall within the row boundaries. The system then selects the window positions that correspond to the maximum total row score and total column score, and uses the locations of the boundaries as the locations of row and column "stripes" that are sized to the spot diameters. The intersections of the row and column stripes define the segments for the respective spots, and the centers of the intersections are the locations for the spot location, or grid, markers. The system may also determine the best spot size, column and row gap sizes and sub-array gap sizes by changing the boundaries accordingly and determining corresponding maximum scores.
|
