Took the output of segmentation as input and then apply svm and baysain for classification. I want exect(in 3*3 matrix) output as that results from svm and Bayesian both side. Took the fs4 in the code as input do apply the svm and bayesian algo on that as calcifications.
TABLE 1: Classification matyrix of wound tissue pixels using Bayesian learning. |
Original |
Granulation |
Predicted pixels Slough |
Necrotic |
Tissue-wise accuracy (%) |
Overall accuracy |
Granulation |
192 |
21 |
9 |
86.48 |
81.15% |
Slough |
67 |
353 |
31 |
78.27 |
Necrotic |
5 |
15 |
74 |
78.72 |
TABLE 2: Wound pixel classification matrix using SVM learning models. |
SVM |
Original |
Granulation |
Predicted pixels Slough |
Necrotic |
Linear kernel |
Granulation |
184 |
35 |
3 |
Slough |
50 |
390 |
11 |
Necrotic |
14 |
34 |
46 |
2nd order polynomial |
Granulation |
182 |
33 |
7 |
Slough |
35 |
400 |
13 |
Necrotic |
5 |
22 |
69 |
3rd order polynomial |
Granulation |
195 |
23 |
4 |
Slough |
31 |
410 |
10 |
Necrotic |
3 |
16 |
75 |
RBF kernel |
Granulation |
184 |
32 |
6 |
Slough |
39 |
401 |
11 |
Necrotic |
4 |
20 |
70 |
Attachment:- fin.rar