Table 4.
Accuracy results for classification of anatomical types for IEEE ISBI 2015 Challenge Dataset 2.7
| Reference | IEEE ISBI 2015 Challenge Dataset 2 | ||||||
|---|---|---|---|---|---|---|---|
| Confusion matrix (proposed) | Diagonal average (proposed, %) | Diagonal average28,33–35 (%) | Diagonal average37 (%) | ||||
| ANB |
Type 1 (%) | Type 2 (%) | Type 3 (%) | 77.31 |
76.64 |
75.83 |
|
| Type 1 | 61.29 | 3.23 | 35.48 | ||||
| Type 2 | 18.52 | 77.78 | 3.70 | ||||
| Type 3 |
7.14 |
0.00 |
92.86 |
||||
| SNB |
Type 1 | Type 2 | Type 3 | 69.81 |
75.24 |
81.92 |
|
| Type 1 | 58.06 | 12.90 | 29.03 | ||||
| Type 2 | 23.08 | 69.23 | 7.69 | ||||
| Type 3 |
16.07 |
1.79 |
82.14 |
||||
| SNA |
Type 1 | Type 2 | Type 3 | 66.72 |
70.24 |
77.97 |
|
| Type 1 | 60.47 | 20.93 | 18.60 | ||||
| Type 2 | 22.50 | 75.00 | 2.50 | ||||
| Type 3 |
29.41 |
5.88 |
64.70 |
||||
| ODI |
Type 1 | Type 2 | Type 3 | 72.28 |
63.71 |
71.26 |
|
| Type 1 | 72.22 | 7.41 | 20.37 | ||||
| Type 2 | 35.00 | 60.00 | 5.00 | ||||
| Type 3 |
11.54 |
3.85 |
84.62 |
||||
| APDI |
Type 1 | Type 2 | Type 3 | 87.18 |
79.93 |
87.25 |
|
| Type 1 | 76.19 | 14.29 | 9.52 | ||||
| Type 2 | 4.55 | 90.91 | 4.55 | ||||
| Type 3 |
5.56 |
0.00 |
94.44 |
||||
| FHI |
Type 1 | Type 2 | Type 3 | 69.16 |
86.74 |
90.90 |
|
| Type 1 | 82. 98 | 2.13 | 14.89 | ||||
| Type 2 | 0.00 | 50.00 | 50.00 | ||||
| Type 3 |
23.53 |
1.96 |
74.51 |
||||
| FMA |
Type 1 | Type 2 | Type 3 | 78.01 |
78.90 |
80.66 |
|
| Type 1 | 60.71 | 35.71 | 3.57 | ||||
| Type 2 | 8.33 | 90.00 | 1.67 | ||||
| Type 3 |
8.33 |
8.33 |
83.33 |
||||
| MW | Type 1 | Type 2 | Type 3 | 77.45 | 77.53 | 82.11 | |
| Type 1 | 78.05 | 9.76 | 12.20 | ||||
| Type 2 | 15.38 | 84.62 | 0.00 | ||||
| Type 3 | 21.21 | 9.09 | 69.70 | ||||