Research Article
Finger Vein Recognition Using Optimal Partitioning Uniform Rotation Invariant LBP Descriptor
Table 3
Block selection using the estimate of the amount of information in each block.
(a) Uniform LBP |
| Block number | Recognition rate | Feature dimension |
| 1 | 0.7500 | | 3 | 0.9787 | | 5 | 0.9920 | | 7 | 0.9973 | | 8 | 0.9987 | | 9 | 0.9987 | | 11 | 0.9987 | | 13 | 0.9987 | | | | | 18 | 0.9987 | |
|
|
(b) Rotation invariant LBP |
| Block number | Recognition rate | Feature dimension |
| 1 | 0.2600 | | 3 | 0.8280 | | 5 | 0.9413 | | 7 | 0.9827 | | 9 | 0.9867 | | 11 | 0.9920 | | 12 | 0.9973 | | 13 | 0.9973 | | | | | 18 | 0.9973 | |
|
|
(c) Uniform rotation invariant LBP |
| Block number | Recognition rate | Feature dimension |
| 1 | 0.2560 | | 3 | 0.8320 | | 5 | 0.9480 | | 7 | 0.9813 | | 9 | 0.9893 | | 11 | 0.9933 | | 12 | 0.9960 | | 13 | 0.9987 | | | | | 18 | 0.9987 | |
|
|
(d) GLCM |
| Block number | Recognition rate | Feature dimension |
| 1 | 0.2467 | | 3 | 0.5867 | | 5 | 0.7653 | | 7 | 0.8893 | | 9 | 0.9187 | | 11 | 0.9440 | | 13 | 0.9600 | | 15 | 0.9707 | | 17 | 0.9800 | | 18 | 0.9800 | |
|
|