Claim Missing Document
Check
Articles

Found 5 Documents
Search

Facial Emotion Recognition Based on Convolutional Neural Network Using FER2013 Dataset Muhammad Al Faris Syabil; Lailan Sofinah Harahap; Muhammad Rafiq Nasution
Jurnal Teknologi informasi dan Ilmu Komputer Vol. 2 No. 1 (2026): Januari 2026
Publisher : Nolsatu Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.65258/jutekom.v2.i1.44

Abstract

Facial emotion recognition is an important research area in computer vision and artificial intelligence, with applications in human–computer interaction, affective computing, and intelligent systems. This study aims to evaluate the performance of a Convolutional Neural Network (CNN) for facial emotion recognition using the FER2013 dataset. The FER2013 dataset consists of grayscale facial images with a resolution of 48×48 pixels and includes seven emotion classes: angry, disgust, fear, happy, neutral, sad, and surprise. Due to its low image resolution and imbalanced class distribution, FER2013 presents significant challenges for emotion classification tasks. An experimental research approach was employed by implementing a baseline CNN architecture composed of convolutional, pooling, and fully connected layers. Image normalization and batch-based data generation were applied during preprocessing. The model was trained using the Adam optimizer with categorical cross-entropy loss, and an early stopping mechanism was utilized to prevent overfitting. Model performance was evaluated using accuracy, precision, recall, F1-score, and confusion matrix analysis. The experimental results show that the proposed CNN model achieved an overall test accuracy of 55.50%. Emotions with distinctive facial features, such as happy and surprise, obtained higher F1-scores, while minority and visually subtle classes, particularly disgust and fear, exhibited lower performance. These findings indicate that a simple CNN architecture can provide reasonable performance on challenging facial emotion datasets while highlighting the impact of class imbalance and limited image resolution. The proposed model can serve as a baseline for further improvements in facial emotion recognition systems.
Identifikasi Pengenalan Tanda Tangan Menggunakan Algoritma Backpropagation dan GLCM (Grey Level Co-Occurrence Matrix) Dea Alya; Lailan Sofinah Harahap; Dodyk Fahlome
Jurnal Sistem Informasi dan Sistem Komputer Vol 11 No 1 (2026): Vol 11 No 1 - 2026
Publisher : STIMIK Bina Bangsa Kendari

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.51717/simkom.v11i1.1215

Abstract

Tanda tangan adalah salah satu biometrik berbasis perilaku yang sering digunakan dalam proses autentikasi. Namun, berbagai jenis tanda tangan menyebabkan proses identifikasi menjadi kompleks, sehingga memerlukan penggunaan teknologi komputer yang baik. Studi ini dibuat sebagai upaya untuk menganalisis tanda tangan yang dimiliki oleh pemilik tanda tangan berdasarkan tekstur citra menggunakan metode ekstraksi Grey Level Co-occurrence Matrix (GLCM) dan klasifikasi Backpropagation. Empat fitur GLCM (contrast, correlation, energy, homogeneity) dihitung dalam empat arah (0°, 45°, 90°, 135°). Data terdiri dari sekitar 80 data dari 4 kelas, dibagi menjadi 20 data uji dan 60 data latih. Pengujian dilakukan lima kali untuk setiap konfigurasi neuron guna memperoleh hasil rata-rata. Hasil penelitian ini menunjukkan bahwa konfigurasi lapisan tersembunyi dengan 50 neuron memberikan kinerja terbaik dengan akurasi rata-rata sekitar 80%. Meningkatkan jumlah neuron cenderung mengurangi akurasi karena kemungkinan overfitting. Meskipun demikian, kombinasi GLCM dan Backpropagation dapat mengidentifikasi tanda tangan dengan cukup baik.
Classification of Herbal Plants Based on Leaf Images Using Gray Level Co-Occurrence Matrix and K-Nearest Neighbor Fahmi Nur Alimsyah Purba; Fathi Athallah Z; Alfin Alfarizi; Lailan Sofinah Harahap
Journal of Artificial Intelligence and Engineering Applications (JAIEA) Vol. 5 No. 3 (2026): June 2026
Publisher : Yayasan Kita Menulis

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59934/jaiea.v5i3.2291

Abstract

Herbal plants have long been used as traditional medicine. However, many people struggle to tell different herbal leaves apart because they look quite similar. This study tries to build a system that can recognize two types of herbal leaves, Moringa and Katuk, simply from their photos. We used GLCM to extract texture features from the leaves, then classified them using KNN. The dataset came from Kaggle, with 480 leaf images in total. Before processing, we cropped the images, resized them to 256x256 pixels, and converted them to grayscale. GLCM features were taken from four angles (0°, 45°, 90°, 135°) and then averaged. This gave us four texture values: contrast, correlation, energy, and homogeneity. We tested KNN with k values from 1 to 15 and five different distance metrics. The best result we got was 94% accuracy, using Manhattan distance with k=1. This system could help everyday people identify medicinal plants more easily without needing lab tests.
Comparative Analysis of Sobel, Prewitt, and Canny Methods in Detecting Object Edges in Betta Fish Images Alfin Alfarizi; Cici El Dirrah Syafitri Simanungkalit; Fahmi Nur Alimsyah Purba; Lailan Sofinah Harahap
Journal of Artificial Intelligence and Engineering Applications (JAIEA) Vol. 5 No. 3 (2026): June 2026
Publisher : Yayasan Kita Menulis

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59934/jaiea.v5i3.2293

Abstract

Edge detection is a crucial stage in digital image processing for recognizing the shape and structure of an object. The application of edge detection to betta fish images presents a unique challenge due to their layered, intricately textured, and often semi-transparent fin morphology. This study aims to analyze and compare the performance of three edge detection algorithms, namely Sobel, Prewitt, and Canny, in extracting shape features from betta fish images. The research methodology involved converting the dataset images into a grayscale format and subsequently implementing the three algorithms using the OpenCV library in the Python programming language. The evaluation was conducted visually by observing the sharpness of the edge lines, object continuity, and the occurrence of noise. The results indicate that the Canny algorithm provides the most optimal performance, as it is capable of detecting the thin edge lines of the fish fins with greater detail and continuity due to its hysteresis thresholding process. Meanwhile, the Sobel and Prewitt methods produced thicker edge lines but were less sensitive to the details of the transparent fins. This study is expected to serve as a reference in selecting the appropriate segmentation method for biological objects with complex morphologies.
Prediksi Curah Hujan Bulanan Di Medan Menggunakan Metode Long Short-Term Memory (LSTM) Dedek; Lailan Sofinah Harahap; Muhammad Rayhans Adrian
Jurnal Ilmu Teknologi Informasi Indonesia Vol. 2 No. 1 (2026): JITIFNA - Januari
Publisher : CV. SINAR HOWUHOWU

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.70134/jitifna.v2i1.969

Abstract

This study aims to predict monthly rainfall in Medan City using the Long Short-Term Memory (LSTM) method. The data utilized in this research comprises monthly rainfall figures and the number of rainy days for the 2015–2023 period, obtained from the Meteorology, Climatology, and Geophysics Agency (BMKG) Region I Medan via official publications of the Central Statistics Agency (BPS) of North Sumatra Province. The pre-processing stage involves data cleaning, normalization, and the construction of a time series dataset using a sliding window structure. The LSTM model was developed with two hidden layers and optimized using the Adam algorithm. Evaluation results indicate that the LSTM model effectively captures seasonal patterns and rainfall trends, as evidenced by a low Root Mean Square Error (RMSE) value. This study is expected to serve as a reference for hydrometeorological disaster mitigation in the Medan region.