Research Article
Obscenity Detection Using Haar-Like Features and Gentle Adaboost Classifier
Table 1
Strength and weakness of different nipple detection methods.
| | Methods | Detection rate (%) | False positive (FP) % | False negative (FN) % |
| | Self-organizing map (SOM) [10] | 65.40 | 0.22 | 34.60 | | Adaboost [20] | 75.64 | 17.40 | 24.40 | | Cascaded Adaboost (haar-cascade) [21] | 90.37 | 7.46 | 4.86 | | Gentle Adaboost with haar-cascade (our approach) | 98.75 | 1.00 | 1.25 | | Gentle Adaboost with train-cascade (our approach) | 84.29 | 22.22 | 15.71 |
|
|
