Article Per Year (5 Year)
p-Index From 2019 - 2024
0.882
P-Index
This Author published in this journals
All Journal International Journal of Electrical and Computer Engineering ComEngApp : Computer Engineering and Applications Journal TEKNIK INFORMATIKA Proceeding of the Electrical Engineering Computer Science and Informatics Annual Research Seminar PROCEEDING IC-ITECHS 2014 Proceeding of the Electrical Engineering Computer Science and Informatics Sriwijaya Journal of Informatics and Applications Jurnal Generic
M. Fachrurrozi .
Computer Science Faculty, Universitas Sriwijaya
Computer Science & IT Education Electrical & Electronics Engineering Engineering

Published : 13 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 13 Documents
Search

 1   2 
Real-Time Occluded Face Identification Using Deep Learning Muhammad Fachrurrozi; Anggina Primanita; Rafly Pakomgan; Abdiansah Abdiansah
JURNAL TEKNIK INFORMATIKA Vol 16, No 1 (2023): JURNAL TEKNIK INFORMATIKA
Publisher : Department of Informatics, Universitas Islam Negeri Syarif Hidayatullah

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15408/jti.v16i1.31211

Abstract

One of the most difficult aspects of face identification is face occlusion. Face occlusion is when anything is placed over the face, for example, a mask. Masks occlude multiple important facial features, like the chin, lips, nose, and facial edges. Face identification becomes challenging when important facial features are occluded. Using one of the deep learning algorithms, YOLOv5, this work tries to identify the face of someone whose face is occluded by a mask in real-time. A special program is being created to test the effectiveness of the YOLOv5 algorithm. 14 people's data were registered, and each person had 150 images used for training, validation, and testing. The images used are regular faces and mask-occluded faces. Nine distinct configurations of epoch and batch sizes were used to train the model. Then, during the testing phase, the best-performing configuration was chosen. Images and real-time input were used for testing. The highest possible accuracy of image identification is 100%, whereas the maximum accuracy of real-time identification is 64%. It was found during the testing that the brightness of the room has an influence on the performance of YOLOv5. Identifying individuals becomes more challenging when there are significant changes in brightness.
Sign Language A-Z Alphabet Introduction American Sign Language using Support Vector Machine Muhammad Rasuandi; Muhammad Fachrurrozi; Anggina primanita
Sriwijaya Journal of Informatics and Applications Vol 4, No 2 (2023)
Publisher : Fakultas Ilmu Komputer Universitas Sriwijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36706/sjia.v4i2.74

Abstract

Deafness is a condition where a person's hearing cannot functionnormally. As a result, these conditions affect ongoing interactions,making it difficult to understand and convey information.Communication problems for the deaf are handled through theintroduction of various forms of sign language, one of which isAmerican Sign Language. Computer Vision-based sign languagerecognition often takes a long time to develop, is less accurate, andcannot be done directly or in real-time. As a result, a solution isneeded to overcome this problem. In the system training process,using the Support Vector Machine method to classify data and testingis carried out using the RBF kernel function with C parameters,namely 10, 50, and 100. The results show that the Support VectorMachine method with a C parameter value of 100 has betterperformance. This is evidenced by the increased accuracy of the RBFC=100 kernel, which is 99%.
Segmentation of Skin Lesions Using Convolutional Neural Networks Firdaus Firdaus; Muhammad Fachrurrozi; Muhammad Naufal Rachmatullah; Dewi Chayanti; Annisa Darmawahyuni; Anggun Islami; Ade Iriani Sapitri; Bambang Tutuko
Computer Engineering and Applications Journal Vol 12 No 1 (2023)
Publisher : Universitas Sriwijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.18495/comengapp.v12i1.466

Abstract

Skin lesions play a crucial role as the initial clinical symptoms of diseases such as chickenpox and melanoma. By employing digital image processing techniques for skin cancer detection, it becomes feasible to diagnose these conditions without the need for physical contact with the skin. However, the automatic analysis of dermoscopy images, which exhibit characteristics like residue (hair and ruler markers), indistinct borders, varying contrast, and variations in shape and color, poses significant challenges. To overcome these difficulties, effective hair removal through segmentation has been explored extensively in the literature. In this study, we present a skin lesion segmentation system developed using the Convolutional Neural Networks (CNNs) method with the U-Net architecture. The model was constructed and evaluated using the HAM10000 Dataset. The results achieved by the best-performing model were outstanding, with a Pixel Accuracy, Intersection over Union (IoU), and F1 Score of 95.89%, 90.37%, and 92.54%, respectively
Co-Authors Abdiansah Abdiansah Ade Iriani Sapitri Anggun Islami Anna Dwi Marjusalinah Annisa Darmawahyuni Auzan Lazuardi Bambang Tutuko Clara Fin Badillah Desty Rodiah Dewi Chayanti Erwin, Erwin Firdaus Firdaus Jerry Satiamy Saputra Junia Erlina Kanda Januar Miraswan Lidya Irfiyani Silaban Mardiana Mardiana Muhammad Naufal Rachmatullah Muhammad Naufal Rachmatullah Muhammad Omar Braddley Muhammad Rasuandi Novi Yusliani Novi Yusliani Osvari Arsalan Primanita, Anggina Rafly Pakomgan Raihan Mufid Setiadi Saparudin Saparudin Saparudin Saparudin Saparudin, Saparudin Winda Agusthia Yoga Yuniadi Yunita Yunita