Fig 1.
Flowchart of improved clustering-based image segmentation method.
Fig 2.
Examples of image preprocessing with (a) source image, (b) contrast enhanced image, (c) gridding image, and (d) coordinates of gridding lines.
Table 1.
Features used in our approach.
Fig 3.
Deficiency examples of different segmentation results.
Table 2.
Description of six real microarray data sets.
Fig 4.
Segmentation results of four methods on microarray images drawn from six data sets.
Fig 5.
Gene expression for images in Fig 4 with four different methods.
Table 3.
Special gene expression comparison of four methods on 6 data sets.
Fig 6.
Scatter plot of two channel intensities for four methods on image drawn from DeRisi dataset.
Fig 7.
Scatter plot of two channel intensities for four methods on image drawn from SMD dataset.
Fig 8.
M-A plot of four methods on image drawn from DeRisi dataset.
Fig 9.
M-A plot of four methods on image drawn from SMD dataset.
Table 4.
The mean, minimum and maximum background-corrected intensities of four methods for two channels on six datasets.
Fig 10.
Segmentation results of various shape spots.
Table 5.
Quantum analysis of the segmentation results showed in Fig 10.
Fig 11.
Examples of segmentation results on simulation images.
Table 6.
Quantitative evaluation for four methods on simulation image.
Table 7.
Computational time comparison of five methods on blocks drawn from six data sets (seconds).