Article Per Year (5 Year)
p-Index From 2021 - 2026
2.932
P-Index
This Author published in this journals
All Journal Seminar Nasional Aplikasi Teknologi Informasi (SNATI) Jurnal Teknologi Informasi dan Ilmu Komputer Indonesia Symposium on Computing Indonesian Journal on Computing (Indo-JC) IJoICT (International Journal on Information and Communication Technology) Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) Technomedia Journal Journal of Electronics, Electromedical Engineering, and Medical Informatics Jurnal Sistem Komputer dan Informatika (JSON) Jurnal Teknik Informatika (JUTIF) Charity : Jurnal Pengabdian Masyarakat Indonesian Community Journal Jurnal Locus Penelitian dan Pengabdian eProceedings of Engineering Community Service Seminar and Community Engagement (COSECANT) Jurnal Pengabdian Kolaborasi dan Inovasi IPTEKS BATIK: Jurnal Pengembangan dan Pengabdian Masyarakat Multikultural LOGIC: Jurnal Penelitian Informatika IJoICT (International Journal on Information and Communication Technology)
Bedy Purnama
Telkom University
Religion Aerospace Engineering Agriculture, Biological Sciences & Forestry Arts Humanities Automotive Engineering Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Economics, Econometrics & Finance Education Electrical & Electronics Engineering Engineering Environmental Science Industrial & Manufacturing Engineering Languange, Linguistic, Communication & Media Mathematics Public Health Social Sciences Other

Published : 34 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 34 Documents
Search

 1   2   3   4 
Pengoptimasian Pengukuran Kepadatan Jalan Raya Dengan Cctv Menggunakan Metode Yolov8 : Optimizing Highway Density Measurement with CCTV Using the Yolov8 Method ​ Gibran, Hilal; Purnama, Bedy; Kosala, Gamma
Technomedia Journal Vol 9 No 1 Juni (2024): TMJ (Technomedia Journal)
Publisher : Pandawan Incorporation, Alphabet Incubator Universitas Raharja

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33050/tmj.v9i1.2216

Abstract

The development and growth of society according to research on the website dataindonesia.id as of December 31, 2022, the number was 126.99 million units by the end of last year. Increasing transportation density has become a serious problem in Indonesia. The relationship between speed and traffic flow (volume) can be used as a guide in determining the mathematical value of road capacity under ideal conditions. The proposed system requires CCTV to run properly. In each input frame, the system will perform data preprocessing to determine the vehicle object that will be segmented in the image. When the frame enters and preprocessing is done, the data will be rezoned to adjust to the system's compatibility. Then the data will go through the image processing stage. Image processing uses RGB color and is converted to grayscale in order to distinguish blobs in the frame. When the blob is detected, the number of objects will be counted and calculated to output the number of vehicles. The results of the number of vehicles will be used for datasets in vehicle density optimization. Motion detection that can be applied to measure highway density with CCTV using YOLO.
Time Series Classification of Badminton Pose using LSTM with Landmark Tracking Purnama, Bedy; Erfianto, Bayu; Wirawan, Ilo Raditio
Journal of Electronics, Electromedical Engineering, and Medical Informatics Vol 7 No 1 (2025): January
Publisher : Department of Electromedical Engineering, POLTEKKES KEMENKES SURABAYA

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35882/jeeemi.v7i1.488

Abstract

Traditional methods of analyzing badminton matches, such as video movement analysis, are time-consuming, prone to errors, and rely heavily on manual annotation. This creates challenges in accurately and efficiently classifying badminton actions and player poses. This paper aims to develop an accurate time series classification method for badminton poses using landmark tracking. The proposed method integrates Long Short-Term Memory (LSTM) networks with landmark tracking to classify badminton poses in a time series, addressing the limitations of traditional video analysis techniques. The dataset consists of 30 respondents performing three distinct activities—lob, smash, and serve—under two conditions: good and bad execution. The approach combines LSTM networks with landmark tracking data, utilizing intra-class variation from a multi-view dataset to enhance pose classification accuracy. The LSTM model achieved high accuracy in classifying badminton poses, successfully detecting serves, lobs, and smashes in real-time with over 90% accuracy. Additionally, the system improved match analysis, achieving 85% accuracy in detection and classification, demonstrating the effectiveness of combining landmark tracking with machine learning for sports analysis. This study underscores the importance of pose estimation in badminton analysis, particularly through landmark tracking, which significantly improves the accuracy of classifying player poses and contributes to the advancement of automated sports analysis.
Application of VGG16 in Automated Detection of Bone Fractures in X-Ray Images Adhyaksa, Resky; Purnama, Bedy
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) Vol 9 No 1 (2025): February 2025
Publisher : Ikatan Ahli Informatika Indonesia (IAII)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/resti.v9i1.6101

Abstract

The purpose of this research is to determine whether or not a deep learning model called VGG16 can automatically identify bone fractures in X-ray pictures. The dataset, sourced from Kaggle, includes 10,522 images of human hand and foot bones, which underwent preprocessing steps such as normalization and resizing to 224x224 pixels to enhance data quality. The study utilizes the VGG16 architecture, pre-trained on ImageNet, as a base model, with transfer learning applied to adapt the model for fracture detection by fine-tuning its weights. This architecture consists of five blocks of convolutional and max-pooling layers to effectively extract and enhance information from the images for precise classification. The training and testing phases utilized an 80:20 split of the data, employing binary cross-entropy as the loss function and the Adam optimizer for efficient weight updates. The model achieved high performance, with an accuracy of 99.25%, precision of 98.62%, recall of 98.88%, and an F1-score of 99.16% over 25 epochs with a batch size of 128. Experimental results indicate that smaller batch sizes generally enhance accuracy and reduce loss values, with batch sizes of 128 and 16 yielding optimal performance. The study's findings underscore the potential of VGG16 in improving diagnostic accuracy and reliability in medical imaging, providing a robust tool for fracture detection. Future research should continue exploring hyperparameter optimization to further enhance model performance while balancing computational efficiency.
Transfer Learning pada Estimasi Pose Hewan Menggunakan YoloV8 dan FineTuning Fauzi, Roki; Purnama, Bedy; Erfianto, Bayu
eProceedings of Engineering Vol. 11 No. 4 (2024): Agustus 2024
Publisher : eProceedings of Engineering

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Kemajuan dalam teknologi pengolahan citra dan kecerdasan buatan telah membuka peluang baru dalam analisis citra, terutama dalam konteks estimasi pose hewan. Penelitian ini bertujuan menggabungkan keunggulan YOLOV8 dalam deteksi objek dengan akurasi estimasi pose hewan melalui pendekatan transfer learning. Dengan melakukan finetuning pada YOLOV8 menggunakan dataset khusus untuk estimasi pose hewan, penelitian ini berupaya meningkatkan kemampuan model dalam mengenali dan menentukan posisi berbagai bagian tubuh hewan dengan lebih tepat. Suksesnya penelitian ini diharapkan dapat memberikan kontribusi pada pengembangan estimasi pose hewan, membuka peluang dalam pengelolaan kesehatan hewan, studi perilaku hewan, dan aplikasi lain yang membutuhkan analisis citra yang kompleks. Namun, penelitian ini memiliki batasan, termasuk fokus eksklusif pada estimasi pose hewan melalui teknik transfer learning dan fine-tuningg. Kata Kunci: Stanford Dog Dataset, YOLOV8, finetuning, transfer learning.
Pembekalan Berpikir Komputasional Untuk Guru-Guru Homeschooling Sahabat Anak Terang Pengajar Anak Special Needs Gunawan, Putu Harry; Pudjoadmojo, Bambang; Rachmawati, Ema; Purnama, Bedy; Sujana, Aprianti Putri; Rudawan, Rikman Aherliwan
Charity : Jurnal Pengabdian Masyarakat Vol. 7 No. 1 (2024): Charity - Jurnal Pengabdian Masyarakat
Publisher : PPM Universitas Telkom

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Sekolah Homeschooling Sahabat Anak Terang adalah layanan sekolah inklusi yang terbuka melayani anak-anak jenjang SD dan SMP melalui pendekatan multisensori. Setiap anak dibuatkan kurikulum individu dalam bentuk Individual Lesson Plan (ILP). Sekolah ini memiliki kurikulum khusus untuk menangani satu anak yang memiliki kebutuhan khusus. Area kurikulum Homeschooling Sahabat Anak Terang meliputi tiga (3) area yaitu Literasi, Matematika, dan Project-based Learning. Kurikulum ini disusun secara terstruktur dan konseptual melalui pengalaman-pengalaman multisensori. Berpikir Komputasional (BK) merupakan konsep berpikir secara Informatika melalui beberapa konsep seperti logika, abstraksi, dekomposisi, algoritma, dan pengenalan pola. Konsep ini sangat penting untuk diberikan pada semua aspek mata pelajaran yang ada di sekolah. Tujuan diadakan kegiatan Pengabdian Kepada Masyarakat (PKM) ini adalah untuk memberikan pengetahuan kepada guru-guru di Homeschooling Sahabat Anak Terang terkait pola berpikir komputasional. Pola berpikir komputasional ini belum sepenuhnya dimengerti oleh guru-guru di sekolah tersebut sehingga mereka sangat tertarik untuk menerapkan konsep ini. Hasil dari kegiatan ini berupa pelatihan dan sosialisasi pengembangan konsep BK dalam beberapa tahapan, seperti pemaparan, latihan soal, permainan dan bedah BK ke mata pelajaran. Berdasarkan umpan balik masyarakat atau guru-guru dalam kegiatan PKM ini, didapatkan sebesar 97,1% peserta dari 35 orang sangat setuju dan setuju bahwasannya kegiatan yang dimaksud sesuai dengan kebutuhan mitra. Selain itu, 100% peserta sangat setuju dan setuju kegiatan PKM dilanjutkan di masa yang akan datang.
Vision Transformer untuk Klasifikasi Kematangan Pisang Pangestu, Arya; Purnama, Bedy; Risnandar, Risnandar
Jurnal Teknologi Informasi dan Ilmu Komputer Vol 11 No 1: Februari 2024
Publisher : Fakultas Ilmu Komputer, Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.25126/jtiik.20241117389

Abstract

Produksi pisang di Indonesia pada tahun 2022 mencapai 9,6 juta ton buah. Metode konvensional yang digunakan untuk menentukan tingkat kematangan pisang masih mengandalkan indera penglihatan manusia dengan memperhatikan perubahan warna kulit pisang. Namun, penentuan tingkat kematangan pisang dengan metode ini memiliki beberapa kekurangan, seperti waktu yang lama, penilaian yang bersifat subjektif dan dapat menghasilkan hasil yang berbeda-beda bagi setiap individu. Oleh karena itu, teknologi computer vision dapat menjadi solusi yang efektif dalam mengklasifikasikan kematangan buah pisang secara otomatis. Penelitian ini menggunakan metodologi Vision Transformer (ViT) untuk mengklasifikasikan tingkat kematangan pada buah pisang, dengan tingkatan yang dibagi menjadi empat kategori, yaitu mentah, setengah matang, matang, dan terlalu matang. Penelitian dilakukan dengan menggunakan lima model ViT yang sudah dilatih sebelumnya atau pre-trained, yaitu ViT-B/16, ViT-B/32, ViT-L/16, ViT-L/32, and ViT-H/14 pada ImageNet-21k dan ImageNet-1k. Kemudian, model ViT tersebut dievaluasi dan dibandingkan dengan model CNN. Evaluasi dilakukan menggunakan metode cross-dataset dengan 5.068 citra pisang yang berbeda dari dataset latih. Hasil evaluasi menunjukkan model ViTL/16-in21k memiliki akurasi tertinggi sebesar 91,61%. Model ViT menunjukkan kemampuan generalisasi yang lebih baik, sementara CNN memiliki ukuran model dan waktu pelatihan yang lebih efisien.
Video Extraction Into PPG Signal To Identify Blood Pressure With XGBoost Method Rahmawan, Adhan Mulya; Purnama, Bedy; Erfianto, Bayu
Indonesian Journal on Computing (Indo-JC) Vol. 9 No. 2 (2024): August, 2024
Publisher : School of Computing, Telkom University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.34818/INDOJC.2024.9.2.942

Abstract

Abstract
Analisis Kemampuan Beta-VAE Pada Dataset Yang Berbeda Bramantya Purbaya; Purnama, Bedy; Edward Ferdian
LOGIC: Jurnal Penelitian Informatika Vol. 3 No. 2 (2025): Desember 2025
Publisher : Universitas Telkom

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.25124/logic.v3i2.9817

Abstract

Data sintetis sudah menjadi beberapa penelitian untuk kasus machine learning, salah satunya adalah menambah data baru dikarenakan kurangnya data yang sudah ada. Tetapi bagaimana untuk menghasilkan dan mengatur berbagai variasi dari distribusi data masukan masih menjadi bahan penelitian. Pada penelitian ini menggunakan salah satu variasi metode Variational Auto Encoder (VAE) untuk menghasilkan data sintetis, yaitu Beta-Variational Auto Encoder (Beta-VAE). VAE sendiri merupakan metode unsupervised learning yang dapat menghasilkan data sintetis, tetapi variasi yang dihasilkan tidak terlalu teratur dibandingkan Beta-VAE. Pada penelitian ini digunakan metode Beta- VAE asli untuk menghasilkan data sintetis yang dilatih dengan empat dataset yang berbeda. Digunakan metrik PSNR, SSIM dan FID score untuk mengevaluasi model Beta-VAE. Dibandingkan setiap model Beta-VAE yang dilatih dengan dataset berbeda dan dilakukan analisis pada setiap model. Hasil dari penelitian didapati model yang dilatih dengan CelebA memiliki hasil terbaik terlihat dari metrik evaluasi.
Detecting Deepfake Videos Using CNN and GRU Methods: Evaluating Performance on the Celeb-DF(v2) Dataset Afandi, Rusdi; Purnama, Bedy
Jurnal Sistem Komputer dan Informatika (JSON) Vol. 7 No. 2 (2025): Desember 2025
Publisher : Universitas Budi Darma

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30865/json.v7i2.9372

Abstract

The development of deep learning technology has allowed the emergence of the phenomenon of deepfakes, which is the manipulation of digital videos that resemble real videos with a high level of realism. These technologies pose serious threats to privacy, digital security, and the spread of false information. As the quality of deepfake videos increases, the detection of this fake content becomes increasingly challenging. This study aims to design and evaluate a deepfake video detection model using a combination of Convolutional Neural Network (CNN) and Gated Recurrent Unit (GRU). CNN is used to extract spatial features from each video frame, while GRU is used to capture the temporal relationships between frames. The dataset used is Celeb-DF(v2), which is a benchmark dataset that contains real videos and high-quality deepfake videos. The CNN-GRU model was trained and tested on the dataset, and its performance was evaluated using accuracy, precision, recall, and F1-score metrics.
Early Fusion of CNN Features for Multimodal Biometric Authentication from ECG and Fingerprint Using MLP, LSTM, GCN, and GAT Priyatama, Muhammad Abdhi; Nugrahadi, Dodon Turianto; Budiman, Irwan; Farmadi, Andi; Faisal, Mohammad Reza; Purnama, Bedy; Adi, Puput Dani Prasetyo; Ngo, Luu Duc
Jurnal Teknik Informatika (Jutif) Vol. 6 No. 6 (2025): JUTIF Volume 6, Number 6, Desember 2025
Publisher : Informatika, Universitas Jenderal Soedirman

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52436/1.jutif.2025.6.6.5299

Abstract

Traditional authentication methods such as PINs and passwords remain vulnerable to theft and hacking, demanding more secure alternatives. Biometric approaches address these weaknesses, yet unimodal systems like fingerprints or facial recognition are still prone to spoofing and environmental disturbances. This study aims to enhance biometric reliability through a multimodal framework integrating electrocardiogram (ECG) signals and fingerprint images. Fingerprint features were extracted using three deep convolutional networks—VGG16, ResNet50, and DenseNet121—while ECG signals were segmented around the first R-peak to produce feature vectors of varying dimensions. Both modalities were fused at the feature level using early fusion and classified with four deep learning algorithms: Multilayer Perceptron (MLP), Long Short-Term Memory (LSTM), Graph Convolutional Network (GCN), and Graph Attention Network (GAT). Experimental results demonstrated that the combination of VGG16 + LSTM and ResNet50 + LSTM achieved the highest identification accuracy of 98.75 %, while DenseNet121 + MLP yielded comparable performance. MLP and LSTM consistently outperformed GCN and GAT, confirming the suitability of sequential and feed-forward models for fused feature embeddings. By employing R-peak-based ECG segmentation and CNN-driven fingerprint features, the proposed system significantly improves classification stability and robustness. This multimodal biometric design strengthens protection against spoofing and impersonation, providing a scalable and secure authentication solution for high-security applications such as digital payments, healthcare, and IoT devices.
Co-Authors Ade Romadhony Adhan Mulya Rahmawan Adhyaksa, Resky Adi, Puput Dani Prasetyo Afandi, Rusdi Agung Toto Wibowo Andi Farmadi Andre Sitompul Angga Rusdinar Aprianti Putri Sujana Bambang Pudjoatmodjo Bambang Pudjotatmodjo Bayu Erfianto Bramantya Purbaya Danu Hary Prakoso Ditari Salsabila E. Dodi Wisaksono Sudiharto Dodon Turianto Nugrahadi Dwi Fitrizal Salim Edward Ferdian Ema Rachmawati Ema Rachmawati Ema Rachmawati Entik Insanudin Farid Hidayat Fat'hah Noor Prawira Fat’hah Noor Prawira Fat’hah Noor Prawira Fauzi, Roki Fazmah Arif Yulianto Febryanti Sthevanie Ferdian, Edward Furqoon, Naufal Sayyid Gamma Kosala Gibran, Hilal Gryaningrum Widi Pangestuti Hafidz Al Djohari Imamul Akhyar Irwan Budiman Ismail Ismail Koredianto Usman Labib, Fahdi Mahmud Dwi Sulistiyo Mahmud Imrona Marliani Harahap Muhammad Arzaki Muhammad Jendro Yuwono Muhammad Jendro Yuwono Muhammad Reza Faisal, Muhammad Reza Muhammad Shafhi Kasyfillah Ngo, Luu Duc Pangestu, Arya Priyatama, Muhammad Abdhi Pudjoadmojo, Bambang Purbaya, Bramantya Putra, Bima Andika Putu Harry Gunawan Rahmawan, Adhan Mulya Reza Dwi Ansari Rian Febrian Umbara Rikman Aherliwan Rudawan Rimba Whidiana Ciptasari Risnandar, Risnandar Rivan Ardyanto Sutoyo Selly Meliana Setyorini Setyorini Sonia Dian Maniswari Tito Prihambodo Tjokorda Agung Budi Wirayuda Wirawan, Ilo Raditio Yuridikta Adha Muslim