Table 1.
Morphology of cancer cell lines used in this study.
Figure 1.
Sample images from different cancer cell line classes.
a) BT-20, b) Focus, c) HepG2, d) MDA-MB-157, e) MV, f) PLC, g) SkHep1, h) T47D.
Table 2.
Names of cancer cell lines used in this study.
Figure 2.
Examples of misclassified images (20×).
Misclassified images are shown in the first column. Examples from their true cell line are given in the second column. Images in the third column show examples of the cell line that the images got misclassified into.
Table 3.
Average classification accuracies (in %) of 10× carcinoma cell line images over 20 runs using SVM with RBF kernel.
Table 4.
Average classification accuracies (in %) of 20× carcinoma cell line images over 20 runs using SVM with RBF kernel.
Table 5.
Average classification accuracies (in %) of 40× carcinoma cell line images over 20 runs using SVM with RBF kernel.
Table 6.
Classification accuracies for SIFT features.
Table 7.
Classification accuracies for variance values only.
Figure 3.
Filterbanks for the dual-tree complex wavelet transform.
Figure 4.
Examples of segmentation into foreground and background.
a) Original image, b) EM Segmentation, c) EM segmentation followed by morphological closing and median filtering.