Research Article
Convolutional Neural Network Architectures to Solve a Problem of Tuberculosis Classification Using X-Ray Images of the Lungs
Table 5
Results of pneumonia classification using Inception v3 & MobileNet v1.
| | Images | Number of training steps | Cross entropy | Validation accuracy | Final testing accuracy |
| | Results of pneumonia classification using Inception v3 | | Without processing | 4000 | 0.156805 | 83.0% () | 74.6% () | | Processed | 4000 | 0.212796 | 78.0% () | 77.0% () | | Processed resized | 4000 | 0.228533 | 76.0% () | 73.1% () |
| | Results of pneumonia classification using MobileNet v1 | | Without processing | 10,000 | 0.026431 | 86.0% () | 74.6% () | | Processed | 10,000 | 0.105245 | 64.0% () | 72.4% () | | Processed resized | 10,000 | 0.238091 | 78.0% () | 75.0% () |
|
|
