Text
SEGMENTASI AKAR PADA CITRA TANAH DENGAN MENGIMPLEMENTASIKAN METODE ENSEMBLE LEARNING MENGGUNAKAN TEKNIK WEIGHTED AVERAGE PADA ARSITEKTUR U-NET DAN DENSENET
The purpose of this study was to obtain an accurate segmentation model in predicting roots using a soil image dataset. This research implements ensemble learning with a weighted average technique on U-Net and DenseNet architectures for root segmentation in soil images. The research stages are data description, data pre-processing, training, testing, and performance evaluation as well as making conclusions. In this study, the results obtained from the performance evaluation of a single architecture, namely accuracy, sensitivity, specificity, precision, F1-Score, and IoU obtained from the U-Net architecture respectively, are 99.69% in this study, 75.75%, 99.76%, 66.63%, 70.87%, and 54.38%. While the results obtained from DenseNet architecture are 99.52%, 71.33%, 99.72%, 66.68%, 67.29%, and 50.71%. The results of the implementation of ensemble learning on U-Net and DenseNet architectures are 99.75%, 92.82%, 99.87%, 92.68%, 92.75%, and 86.49%, respectively. Based on the results obtained, it can be concluded that the implementation of ensemble learning on U-Net and DenseNet architectures can segment the roots of the soil image very well. The implementation of ensemble learning can also improve architectural performance and overcome the problem of overfitting in single architecture segmentation.
Inventory Code | Barcode | Call Number | Location | Status |
---|---|---|---|---|
2207005183 | T84725 | T847252022 | Central Library (Referens) | Available but not for loan - Not for Loan |
No other version available