Please use this identifier to cite or link to this item:
https://repositori.uma.ac.id/handle/123456789/29708Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor.advisor | Muhathir | - |
| dc.contributor.author | Amelia, Dilla | - |
| dc.date.accessioned | 2026-04-10T04:42:20Z | - |
| dc.date.available | 2026-04-10T04:42:20Z | - |
| dc.date.issued | 2025-09 | - |
| dc.identifier.uri | https://repositori.uma.ac.id/handle/123456789/29708 | - |
| dc.description | 69 Halaman | en_US |
| dc.description.abstract | Penelitian ini dilakukan dengan tujuan meningkatkan ketepatan dalam mengklasifikasikan penyakit pada daun anggur menggunakan pendekatan Deep Learning dengan arsitektur DenseNet169 yang dipadukan dengan CBAM (Convolutional Block Attention Module). Arsitektur DenseNet169 dimanfaatkan untuk mengekstraksi fitur visual penting dari citra, sedangkan CBAM berfungsi untuk memperkuat perhatian model terhadap informasi spasial dan kanal yang paling relevan. Pelatihan model dilakukan secara konsisten selama 20 epoch dan dievaluasi menggunakan sejumlah metrik, yaitu akurasi, presisi, recall, F1-score, dan ROC-AUC. Hasil pengujian menunjukkan bahwa kombinasi model DenseNet169 dan CBAM menghasilkan akurasi sebesar 99,50%, serta presisi, recall, dan F1-score yang sama tinggi. Nilai ROC-AUC yang diperoleh mencapai 99,96%, menandakan performa klasifikasi yang sangat optimal. Jika dibandingkan dengan penelitian sebelumnya yang menggunakan pendekatan serupa, model ini menunjukkan kinerja yang lebih baik. Temuan ini menunjukkan bahwa integrasi CBAM dalam arsitektur konvolusional seperti DenseNet169 berkontribusi positif dalam pengembangan sistem deteksi penyakit tanaman secara otomatis. Ke depan, model ini berpotensi diterapkan dalam sistem monitoring berbasis drone secara real-time, guna mendukung pengambilan keputusan yang lebih tepat dalam sektor pertanian. This study aims to improve the accuracy of grape leaf disease classification by employing a deep learning approach using the DenseNet169 architecture integrated with the Convolutional Block Attention Module (CBAM). DenseNet169 is utilized to extract important visual features from images, while CBAM enhances the model's focus on the most relevant spatial and channel information. The model was trained consistently for 20 epochs and evaluated using several performance metrics, including accuracy, precision, recall, F1-score, and ROC-AUC. The experimental results indicate that the combination of DenseNet169 and CBAM achieved an accuracy of 99.50%, with precision, recall, and F1-score reaching the same value. The ROC-AUC score reached 99.96%, reflecting an excellent classification performance. Compared to previous studies that applied similar approaches, this model demonstrated superior performance. These findings suggest that integrating CBAM into a convolutional architecture like DenseNet169 contributes positively to the development of automated plant disease detection systems. In the future, this model holds potential for real-time implementation in drone-based monitoring systems to support more effective decision-making in the agricultural sector. | en_US |
| dc.language.iso | id | en_US |
| dc.publisher | Universitas Medan Area | en_US |
| dc.relation.ispartofseries | NPM;218160037 | - |
| dc.subject | DenseNet169 | en_US |
| dc.subject | CBAM | en_US |
| dc.subject | Deep Learning | en_US |
| dc.subject | Image Classification | en_US |
| dc.subject | Klasifikasi Citra | en_US |
| dc.subject | Penyakit Daun Aggur | en_US |
| dc.title | Peningkatan Klasifikasi Penyakit Daun Anggur Menggunakan Denseneti69 dengan Mekanisme Perhatian CBAM | en_US |
| dc.title.alternative | Improved Classification of Grape Leaf Diseases Using Denseneti69 with CBAM Attention Mechanism | en_US |
| dc.type | Thesis | en_US |
| Appears in Collections: | SP - Informatic Engineering | |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 218160037 - Dilla Amelia - Chapter IV.pdf Restricted Access | Chapter IV | 2.35 MB | Adobe PDF | View/Open Request a copy |
| 218160037 - Dilla Amelia - Fulltext.pdf | Cover, Abstract, Chapter I, II, III, V, Bibliography | 1.63 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.