SYAFIQ, FREEDAN RAFII (2026) PERANCANGAN DAN IMPLEMENTASI WEBSITE EYESHIELD UNTUKDETEKSI DINI DAN EDUKASI KESEHATAN MATA. S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
01 COVER.pdf Download (2MB) | Preview |
|
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (BAB I)
02 BAB 1.pdf Restricted to Registered users only Download (48kB) |
|
|
Text (BAB II)
03 BAB 2.pdf Restricted to Registered users only Download (154kB) |
|
|
Text (BAB III)
04 BAB 3.pdf Restricted to Registered users only Download (889kB) |
|
|
Text (BAB IV)
05 BAB 4.pdf Restricted to Registered users only Download (369kB) |
|
|
Text (BAB V)
06 BAB 5.pdf Restricted to Registered users only Download (29kB) |
|
|
Text (DAFTAR PUSTAKA)
07 DAFTAR PUSTAKA.pdf Restricted to Registered users only Download (167kB) |
|
|
Text (LAMPIRAN)
08 LAMPIRAN.pdf Restricted to Registered users only Download (533kB) |
Abstract
Eye diseases may cause visual impairment and blindness when they are not detected early. This study develops EyeShield, a web-based system integrating an EfficientNet B0 Convolutional Neural Network and an IndoBERT chatbot. The CNN classifies five classes: Cataract, Conjunctivitis, Eyelid, Normal, and Uveitis. Images are validated, converted to RGB, resized to 224 x 224 pixels, and normalized before inference. The chatbot classifies seven intents and retrieves answers from a Q&A knowledge base using cosine similarity. The AI backend is built with FastAPI and PyTorch, while the web application uses Laravel, Bootstrap, and MySQL connected through a REST API. Evaluation uses Black Box Testing and the Precision, Recall, and F1-Score metrics. All eight documented API test scenarios were successful. The classification report shows a macro-average precision of 0.90, a macro-average recall of 0.898, and a macro-average F1-score of 0.898. The macro-average recall represents a balanced accuracy of 89.8%. The Normal class achieves the highest F1-score of 0.95, whereas the Eyelid class achieves the lowest F1-score of 0.86. EyeShield can support preliminary screening and eye-health education, but it does not replace professional medical diagnosis. Keyword: Early Eye Disease Detection, EfficientNet B0, Convolutional Neural Network, IndoBERT, Chatbot, Non-Fundus Eye Image, FastAPI, Laravel, REST API. Penyakit mata dapat menyebabkan gangguan penglihatan hingga kebutaan apabila tidak dideteksi sejak dini. Penelitian ini mengembangkan website EyeShield yang mengintegrasikan Convolutional Neural Network berarsitektur EfficientNet B0 dan chatbot IndoBERT. Model CNN mengklasifikasikan lima kelas, yaitu Cataract, Conjunctivitis, Eyelid, Normal, dan Uveitis. Citra divalidasi, dikonversi ke RGB, diubah menjadi 224 x 224 piksel, dan dinormalisasi sebelum proses inferensi. Chat bot mengklasifikasikan tujuh intent dan mengambil jawaban dari basis pengetahuan Q&A menggunakan cosine similarity. Backend AI dibangun dengan FastAPI dan PyTorch, sedangkan aplikasi web menggunakan Laravel, Bootstrap, dan MySQL yang terhubung melalui REST API. Pengujian dilakukan menggunakan Black Box Testing serta metrik Precision, Recall, dan F1-Score. Seluruh delapan skenario pen gujian API yang didokumentasikan berhasil. Classification report menunjukkan macro-average precision 0,90, macro-average recall 0,898, dan macro-average F1 score 0,898. Model memperoleh macro-average recall sebesar 89,8%. Kelas Normal memperoleh F1-score tertinggi sebesar 0,95, sedangkan kelas Eyelid memperoleh F1-score terendah sebesar 0,86. EyeShield dapat mendukung skrining awal dan edukasi kesehatan mata, tetapi tidak menggantikan diagnosis tenaga medis.Android. Kata kunci: Deteksi Dini Penyakit Mata, EfficientNet B0, Convolutional Neural Network, IndoBERT, Chatbot, Citra Mata Non-Fundus, FastAPI, Laravel, REST API.
Actions (login required)
 |
View Item |
