Claim Missing Document
Check
Articles

Found 6 Documents
Search

Game Edukasi Siswa SLB Untuk Meningkatkan Literasi Matematika Berbasis Android Hasanuddin, Muhammad Hasrul; Indra Syahyadi, Asep; Darmatasia, Darmatasia; Ridwang, Ridwang
Jurnal INSYPRO (Information System and Processing) Vol 8 No 2 (2023)
Publisher : Prodi Sistem Informasi UIN Alauddin

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24252/insypro.v8i2.41134

Abstract

The research aims to develop and test the effectiveness of an Android-based educational game designed specifically for students of the Extraordinary School (SLB) in improving their mathematical literacy. Mathematical literacy is an essential skill that supports an individual's ability to understand, analyze, and apply mathematical concepts in everyday life. However, SLB students often face the challenge of acquiring a sufficient understanding of mathematics due to various barriers in their learning. This research methodology covers the stages of development of educational games designed in accordance with the characteristics of SLB students and the principles of effective mathematical learning. After the game was developed, the study involved a group of SLB students in a field experiment using a quasi-experimental approach with the control group. During a certain period, students from the experimental group participated in learning sessions using educational games, while students from control groups received conventional mathematical learning. The results of this study are expected to provide insight into the extent to which this Android-based educational game is effective in improving the mathematical literacy of SLB students. Furthermore, the research also has the potential to provide valuable input for further development of similar educational games that can be used as an inclusive learning tool for students with special educational needs. Increased mathematical literacy among SLB students can provide long-term benefits in preparing them to face the demands of mathematics in their later lives
ANALISIS PERFORMA CONVOLUTIONAL NEURAL NETWORK DENGAN HYPERPARAMETER TUNING DALAM MENDETEKSI GAMBAR DEEPFAKE Darmatasia, Darmatasia; Ramli, Abdur Rahman; Salsabila, Azizah; Adiba, Fhatiah
Jurnal INSYPRO (Information System and Processing) Vol 9 No 2 (2024)
Publisher : Prodi Sistem Informasi UIN Alauddin

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24252/insypro.v9i2.51928

Abstract

This research analyzes the performance of Convolutional Neural Network (CNN) in detecting deepfake images with a focus on hyperparameter tuning. The study consists of two classes: fake images and real images, with each class containing 5000 data samples. Hyperparameter tuning is conducted using the Keras-tuner library, a framework used for automatic hyperparameter tuning on models built with Keras, eliminating the need for manual trial and error tuning. The hyperparameter search strategy employed is random search. The results of the study indicate that hyperparameter tuning significantly improves the model's detection accuracy. Various experiments were conducted to evaluate the impact of hyperparameter settings, such as the number and size of filters, learning rate, and optimizer. Analysis of different optimizers showed significant variations in performance, with Adam Optimizer achieving the highest accuracy of 83% using a combination of 32 filters sized 3x3 in the first layer and 128 filters sized 5x5 in the second layer. RMSProp and AdamW each achieved 82% accuracy, SGD Optimizer achieved 75% accuracy, while Adadelta Optimizer achieved 71% accuracy. The findings of this study affirm that the selection of optimizer and appropriate hyperparameter settings have a significant impact on the model's performance in detecting patterns in the data. This study also emphasizes the importance of optimizing filters and sizes in each layer to enhance model accuracy.
PENGENALAN SISTEM ISYARAT BAHASA INDONESIA (SIBI) MENGGUNAKAN GRADIENT-CONVOLUTIONAL NEURAL NETWORK DARMATASIA, DARMATASIA
Jurnal INSTEK (Informatika Sains dan Teknologi) Vol 6 No 1 (2021): APRIL
Publisher : Department of Informatics Engineering, Faculty of Science and Technology, Universitas Islam Negeri Alauddin, Makassar, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (528.563 KB) | DOI: 10.24252/instek.v6i1.18637

Abstract

Penelitian ini bertujuan untuk melakukan pengenalan alfabet pada Sistem Isyarat Bahasa Indonesia (SIBI). Penelitian ini memiliki dua kontribusi utama,  Pertama dilakukan pengumpulan dataset alfabet SIBI. Kedua, pengenalan alfabet SIBI menggunakan algoritma Convolutional Neural Network (CNN). Pada penelitian ini, citra masukan berupa alfabet bahasa isyarat pada lapisan input diberikan filter gradient agar bentuk objek menjadi lebih jelas. Hasil penelitian menunjukkan bahwa pemberian filter pada lapisan input dapat meningkatkan akurasi pengenalan yaitu sekitar 85%. Citra masukan yang tidak difilter hanya memperoleh akurasi sebesar 25%. Akurasi terbaik yang diperoleh yaitu 98% dengan meningkatkan jumlah iterasi. Metode yang diusulkan juga diuji menggunakan tiga benchmark dataset. Hasil pengujian menunjukkan  bahwa metode yang diusulkan dapat meningkatkan akurasi pengenalan pada benchmark dataset yang memiliki background yang kompleks.Kata Kunci: Convolutional Neural Network; Gradient; Sistem Isyarat Bahasa Indonesia 
DETEKSI PENGGUNAAN MASKER MENGGUNAKAN XCEPTION TRANSFER LEARNING DARMATASIA, DARMATASIA
Jurnal INSTEK (Informatika Sains dan Teknologi) Vol 5 No 2 (2020): OCTOBER
Publisher : Department of Informatics Engineering, Faculty of Science and Technology, Universitas Islam Negeri Alauddin, Makassar, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (420.077 KB) | DOI: 10.24252/instek.v5i2.20132

Abstract

Covid-19 telah membawa dampak dalam berbagai aspek kehidupan. Hal tersebut menjadi landasan bagi pemerintah untuk mengambil langkah penerapan protokol kesehatan dalam aktivitas yang dilakukan di luar rumah. Salah satu bagian dari protokol kesehatan adalah kewajiban menggunakan masker pada saat keluar rumah. Saat ini, masih banyak masyarakat yang tidak menggunakan masker saat beraktivitas di luar rumah. Selain itu, terdapat beberapa kalangan yang menggunakan masker namun tidak sesuai dengan standar. Pada penelitian ini akan dilakukan deteksi penggunaan masker dengan menggunakan pendekatan deep learning. Metode yang digunakan yaitu Xception dengan transfer learning. Model yang dikembangkan dapat mendeteksi tiga tipe penggunaan masker yaitu penggunaan masker sesuai dengan standar, penggunaan masker yang tidak sesuai dengan standar, dan tidak menggunakan masker sama sekali. Model yang telah dilatih memperoleh tingkat akurasi sebesar 97%.  Hasil penelitian ini diharapkan dapat dintegrasikan dengan perangkat lain untuk pengembangan sistem deteksi penggunaan masker secara menyeluruh.  Kata Kunci: Masker; Transfer Learning; Xception; 
IMPLEMENTASI CONVOLUTIONAL NEURAL NETWORK UNTUK KLASIFIKASI TANAMAN RIMPANG SECARA VIRTUAL DARMATASIA, DARMATASIA; A. MUHAMMAD SYAFAR
Jurnal INSTEK (Informatika Sains dan Teknologi) Vol 8 No 1 (2023): APRIL
Publisher : Department of Informatics Engineering, Faculty of Science and Technology, Universitas Islam Negeri Alauddin, Makassar, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24252/instek.v8i1.37255

Abstract

Penelitian ini mengimplementasikan Convolutional Neural Network untuk mengklasifikasi jenis tanaman rimpang yang memiliki bentuk dan warna yang hampir sama. Jenis tanaman rimpang yang diklasifikasi yaitu Kunyit, Jahe, Laos, Kencur, dan Kunci. Dalam penelitian ini terdapat 3 arsitektur Convolutional Neural Network yang digunakan yaitu MobileNet, InceptionV3, dan VGGNet. Hasil Penelitian menunjukkan bahwa arsitektur MobileNet dan InceptionV3 memperoleh akurasi yang sama yaitu sebesar 98%. Meskipun demikian, arsitektur MobileNet memiliki waktu komputasi yang lebih cepat. Adapun arsitektur VGG19 memperoleh akurasi sebesar 88%. Hasil penelitian menunjukkan bahwa metode yang digunakan dapat membedakan berbagai jenis tanaman rimpang dengan cukup baik meskipun dengan jumlah data yang sedikit. Penelitian ini diharapkan dapat memudahkan masyarakat atau orang awam yang masih sulit dalam membedakan tanaman rimpang.
Deteksi Penyakit pada Daun Tomat Menggunakan Kombinasi Ekstraksi Fitur Colors Moments dan Grey Level Co-Occurrence Matrix (GLCM) Syarif , Ririn Suharni; Akbar , Muhammad Nur; Darmatasia, Darmatasia
Journal Software, Hardware and Information Technology Vol 5 No 2 (2025)
Publisher : Jurusan Sistem Informasi Universitas Islam Negeri Alauddin Makassar

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24252/shift.v5i2.214

Abstract

Tomato is one of the leading horticultural crops widely cultivated by farmers in Indonesia. In addition to its high economic value, tomatoes are rich in nutrients beneficial to human health, such as vitamin C, lycopene, and other antioxidants. However, tomato productivity is highly vulnerable to decline due to various diseases, particularly those affecting the leaves. These diseases not only reduce the quality of the harvest but also significantly threaten production quantity. Therefore, early detection of leaf diseases in tomato plants is essential to help farmers, especially novice farmers, take timely and appropriate treatment actions. This study aims to develop a digital image-based detection system for tomato leaf diseases using feature extraction methods and classification algorithms. In the image pre-processing and feature extraction stages, the Color Moments algorithm is used to capture color information, while the Gray Level Co-occurrence Matrix (GLCM) represents leaf texture. The classification process is carried out using the Random Forest algorithm. The dataset used in this study was obtained from Kaggle, consisting of 5,451 images of tomato leaves categorized into six classes: Leaf Spot, Leaf Mold, Septoria Leaf Spot, Mosaic Virus, Bacterial Spot, and Healthy Leaf. Test results show that the developed model achieved an accuracy of 90%. These findings indicate that the system can detect tomato leaf diseases with a relatively high level of accuracy. The system is expected to assist farmers, especially beginners, in identifying plant diseases more quickly and accurately, thereby improving treatment efficiency and increasing crop yields.