Abstract
The palmprint is a part of the human body that has unique and detailed characteristics of the pattern of palm lines, such as the length and width of the palm (geometric features), principal lines, and wrinkle lines. It began to be developed as a tool for recognize a person. The palmprint dataset used comes from Kaggle, namely BMPD. The palmprint images in this dataset were taken in 2 sessions. In the first session, there was not much variation in rotation compared to the second session. This research uses Convolutional Neural Network (CNN) models with Xception, VGG16, ResNet50, MobileNet, and EfficientNetB0 architectures to see the best performance. The results of this research showed that the MobileNet model had the best performance with an accuracy of 96.6% and a loss of 14.3%. For Precision results of 94%, Recall 96%, and F1-Score 94%. Meanwhile, Xception obtained an accuracy of 88.3% and a loss of 52.9%, VGG16 70.8% and a loss of 109.8%, ResNet50 5.8% and a loss of 307.9%, and EfficientNetB0 3.3% and a loss of 340.1%.
Concepts :
SDGs
Citations by Year
| Year | Count |
|---|---|
| 2024 | 0 |