18.11.2014 Views

Pit Pattern Classification in Colonoscopy using Wavelets - WaveLab

Pit Pattern Classification in Colonoscopy using Wavelets - WaveLab

Pit Pattern Classification in Colonoscopy using Wavelets - WaveLab

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

5 Results<br />

(a) 2 classes<br />

(b) 6 classes<br />

Figure 5.13: Distance matrices for the two classes and the six classes case (<strong>Pit</strong> pattern images)<br />

Class 1 2 3 4 5 6<br />

1 96 4 0 0 0 0<br />

2 7 93 0 0 0 0<br />

3 0 0 77 9 9 4<br />

4 0 0 10 64 14 12<br />

5 0 0 4 19 52 24<br />

6 0 0 0 8 27 64<br />

Table 5.18: Result distribution matrix for CC for 6 classes (Outex images)<br />

result distribution matrix for the six classes case is quite different compared to the result<br />

distribution matrices for the previous methods, as we notice from table 5.18. We can see<br />

that while the first two classes get classified fairly well, the classification rate is gett<strong>in</strong>g<br />

lower for the classes 3 to 6, with the lowest classification result of 52% only for class 4.<br />

Figure 5.14 shows the distance matrices for the two classes and the six classes case. In<br />

contrast to the distance matrices presented above for the pit pattern images, these matrices<br />

are similar to the example distance matrices presented <strong>in</strong> figure 4.5. Especially for the distance<br />

matrix for the two classes case shown <strong>in</strong> figure 5.14(a) we are able to clearly identify<br />

a typical structure for low <strong>in</strong>tra-class distances and high <strong>in</strong>ter-class distances.<br />

88

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!