Optimization of VGG-16 Accuracy for Fingerprint Pattern Imager Classification
Fingerprint is a unique biometric identity commonly used as evidence in court. However, its quality can decline due to external factors such as uneven surfaces, weather conditions, or distortion. The dataset used in this study is FVC2000. Convolutional Neural Networks (CNN) were applied for fingerpr...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
LPPM ISB Atma Luhur
2025-01-01
|
| Series: | Jurnal Sisfokom |
| Subjects: | |
| Online Access: | https://jurnal.atmaluhur.ac.id/index.php/sisfokom/article/view/2317 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Fingerprint is a unique biometric identity commonly used as evidence in court. However, its quality can decline due to external factors such as uneven surfaces, weather conditions, or distortion. The dataset used in this study is FVC2000. Convolutional Neural Networks (CNN) were applied for fingerprint image enhancement and classification, focusing on patterns such as whorl, arch, radial loop, ulnar loop, and twinted loop. This research optimized the VGG-16 model by adding several hyperparameters. The results showed the highest accuracy of 100% on the testing data with a learning rate of 0.0001, using 50 epochs and a training-to-validation data split ratio of 80%:10% from a total of 400 fingerprint image pattern data. These findings demonstrate that the VGG-16 model successfully classified fingerprint images with optimal performance, contributing significantly to the development of CNN-based fingerprint classification systems. |
|---|---|
| ISSN: | 2301-7988 2581-0588 |