Purboningrum, Fadhila and Prof. Dr.Eng. Fitri Utaminingrum, S.T., M.T and Fais Al Huda, S.Kom., M.Kom. (2024) Sistem Deteksi Dini Autism Spectrum Disorder (ASD) Berbasis Face Recognition Menggunakan Metode Transfer Learning Resnet50. Sarjana thesis, Universitas Brawijaya.
Abstract
Autism Spectrum Disorder (ASD) merupakan suatu kondisi neurologis yang memengaruhi perkembangan anak. ASD memiliki prevalensi global 1 dari 100 anak. Sifat dan gejala ASD yang beragam, seperti kesulitan berkomunikasi, berinteraksi sosial, dan perilaku repetitif menjadi tantangan tersendiri dalam proses diagnosis yang akurat. Karakteristik wajah, diidentifikasi sebagai kunci potensial, menjadi dasar untuk pengembangan sistem deteksi dini menggunakan face recognition dan deep learning. Penelitian ini mengusulkan sistem deteksi dini ASD dengan memanfaatkan Android dan metode transfer learning arsitektur model ResNet50. Sistem ini diimplementasikan dalam bentuk aplikasi Android dengan arsitektur model ResNet50 yang telah dilatih pada dataset ImageNet. Penggunaan teknologi Android dipilih untuk aksesibilitas yang lebih luas, dengan 67% penduduk Indonesia sebagai pengguna smartphone. Hasil penelitian menunjukkan bahwa nilai epoch optimal untuk model adalah 51 dengan nilai accuracy sebesar 91%, nilai precision 92%, serta nilai recall dan f1-score masingmasing sebesar 90%. Selain itu, berdasarkan pengujian dan analisis terhadap hasil pengujian didapatkan akurasi luaran aplikasi sebesar 100% pada kelas autistic dan 60% pada kelas non-autistic dengan rata-rata waktu komputasi 2,827 detik. Saran melibatkan penambahan variasi dataset, validasi data dengan pakar, dan optimasi sisi client Android untuk meningkatkan kecepatan tampilan hasil. Penelitian ini memberikan landasan untuk mendemonstrasikan potensi face recognition dan model deep learning dalam membantu proses diagnosis Autism Spectrum Disorder (ASD) dengan akurasi tinggi, serta untuk mendukung intervensi dini dan meningkatkan kesadaran terhadap ASD.
English Abstract
Autism Spectrum Disorder (ASD) is a neurological condition that affects children's development. ASD has a global prevalence of 1 in 100 children. The diverse nature and symptoms of ASD, such as difficulties in communication, social interaction, and repetitive behaviors, pose a challenge for accurate diagnosis. Facial characteristics, identified as potential keys, form the basis for the development of an early detection system using face recognition and deep learning. This research proposes an ASD early detection system utilizing Android and the ResNet50 model architecture transfer learning method. The system is implemented in the form of an Android application with ResNet50 model architecture that has been trained on ImageNet dataset. The use of Android technology was chosen for wider accessibility, with 67% of Indonesia's population as smartphone users. The results show that the optimal epoch value for the model is 51 with an accuracy value of 91%, a precision value of 92%, and a recall and f1-score value of 90% each. In addition, based on testing and analysis of the test results, the accuracy of the application output is 100% in the autistic class and 60% in the non-autistic class with an average computing time of 2.827 seconds. Suggestions involve adding dataset variations, validating data with experts, and optimizing the Android client side to increase the speed of displaying results. This research provides a foundation to demonstrate the potential of face recognition and the deep learning model in assisting the diagnosis process of Autism Spectrum Disorder (ASD) with high accuracy, as well as to support early intervention and raise awareness of ASD.
| Item Type: | Thesis (Sarjana) |
|---|---|
| Identification Number: | 0524150285 |
| Uncontrolled Keywords: | Aplikasi Mobile Android, Autism Spectrum Disorder (ASD), Face Recognition, Karakteristik Wajah, ResNet50, Transfer Learning-Android Mobile Application, Autism Spectrum Disorder (ASD), Face Recognition, Face Characteristics, ResNet50, Transfer Learning. |
| Divisions: | Fakultas Ilmu Komputer > Teknik Komputer |
| Depositing User: | Sugeng Moelyono |
| Date Deposited: | 03 Jun 2024 07:53 |
| Last Modified: | 03 Jun 2024 07:53 |
| URI: | http://repository.ub.ac.id/id/eprint/220080 |
![[thumbnail of DALAM MASA EMBARGO]](http://repository.ub.ac.id/style/images/fileicons/text.png) |
Text (DALAM MASA EMBARGO)
Fadhila Purboningrum.pdf Restricted to Registered users only Download (5MB) |
Actions (login required)
 |
View Item |