Research Article
Lung Cancer Classification Employing Proposed Real Coded Genetic Algorithm Based Radial Basis Function Neural Network Classifier
Table 7
Accuracy results amongst various classifiers on combining the healthy and cancer affected lung images.
| | Datasets (Combined for Classification) | Image Types | BPN | SVM | ELM | GWO – ELM | SRGWO – ELM | GA – RBFNN | Propose
RCGA– RBFNN
Classifier |
| Lung Image Database Consortium (LIDC)
(152 cases, 48 healthy and 104 cancer affected) | Healthy | 87.8% | 90.6% | 81.2% | 93.2 % | 96.1% | 94.5% | 98.3% | | Well circumscribed nodules | 84.7% | 86.1% | 80.4% | 92.8% | 94.9% | 94.1% | 96.9% | | Vascularized nodules | 80.8% | 85.1% | 79.4% | 91.6% | 95.3% | 96.9% | 98.0% | | Juxtapleural nodules | 84.9% | 88.3% | 75.4% | 89.8% | 91.2% | 90.6% | 95.3% | | Pleural-tail nodules | 78.2% | 75.4% | 70.8% | 86.5% | 89.2% | 88.0% | 91.6% |
| Accura Diagnostics Centre & PSGIMSR
(141 cases, 90 healthy and 51 cancer affected) | Healthy | 90.6% | 94.5% | 87.9% | 96.5% | 98.7% | 98.0% | 99.1% | | Well circumscribed nodules | 86.4% | 90.8% | 84.2% | 93.7% | 97.4% | 96.8% | 98.8% | | Vascularized nodules | 91.8% | 93.2% | 89.0% | 95.1% | 98.0% | 95.7% | 99.6% | | Juxtapleural nodules | 91.6% | 92.4% | 88.0% | 94.9% | 96.2% | 96.0% | 98.7% | | Pleural-tail nodules | 86.4% | 89.7% | 81.6% | 94.3% | 97.8% | 95.9% | 98.9% |
|
|
