Image Processing Reference
In-Depth Information
Ta b l e 9 . 3
Comparison of BCA and BMSRM
Method
Time (s) TPR(%) FPR(%)
BCA
2.63
97.75
1.26
BMSRM
30.28
90.68
8.49
BCA algorithm. BMSRM is easy to fail in some important tissues and wrongly la-
bel background blocks as object (highlighted with red color). Here, an individual
tooth is lost and many holes occur in surface, that dues to BMSRM only considers
local neighborhood similarity while SSL based BCA can achieve a global minimum
error. To get the quantitatively comparison, desired objects are manually sculptured
as ground truth. Table 9.3 lists TPR and FPR comparison between these two meth-
ods. The numeral results imply that BCA achieves the better segmentation perfor-
mance. BCA is also performed in numbers of CBCT sequences and get the similar
results. This validation with low SNR (Signal Noise Ratio) data demonstrates that
by grouping the similar pixels into homogeneous blocks, mean shift initial classifi-
cation improve the robustness of BCA to noises and small pixel variations.
Commonly, the proposed BCA method is not sensitive to initial classification,
this fact makes it is more robust and efficient in volume data segmentation tasks.
Although BCA has been greatly reducing the requirements of user interaction, but
more manual markers are still needed in region with complex structures. Addition-
ally, dental shape and arrangement priors should be considered to increase the pre-
cision.
References
1. Boykov, Y., Kolmogorov, V.: An experimental comparison of min-cut/max-flow algo-
rithms for energy minimization in vision. IEEE Transactions on Pattern Analysis and
Machine Intelligence 26(9) (2004)
2. Boykov, Y., Veksler, O., Zabih, R.: Fast approximate energy minimization via graph cuts.
IEEE Transactions on Pattern Analysis and Machine Intelligence 23(11), 1222-1239
(2001)
3. Boykov, Y.Y., Jolly, M.P.: Interactive graph cuts for optimal boundary and region seg-
mentation of objects in ND images. In: Proceedings of the Eighth IEEE International
Conference on Computer Vision (ICCV 2001), vol. 1, pp. 105-112. IEEE (2001)
4. Burt, P.J., Adelson, E.H.: The Laplacian pyramid as a compact image code. IEEE Trans.
on Communications COM-31(4), 532-540 (1983)
5. Comaniciu, D., Meer, P.: Mean shift: A robust approach toward feature space analysis.
IEEE Transactions on Pattern Analysis and Machine Intelligence 24(5), 603-619 (2002)
6. Comaniciu, D., Ramesh, V., Meer, P.: Kernel-based object tracking. IEEE Transactions
on Pattern Analysis and Machine Intelligence 25(5), 564-577 (2003)
7. Dehmeshki, J., Amin, D., Valdivieso, M., Ye, X.: Segmentation of pulmonary nodules
in thoracic CT scans: A region growing approach. IEEE Transactions on Medical Imag-
ing 27(4) (April 2008)
Search WWH ::
Custom Search