Claim Missing Document
Check
Articles

Found 10 Documents
Search

Multi-scale input reconstruction network and one-stage instance segmentation for enhancing heart defect prediction rate Sutarno, Sutarno; Nurmaini, Siti; Sapitri, Ade Iriani; Rachmatullah, Muhammad Naufal; Tutuko, Bambang; Darmawahyuni, Annisa; Firdaus, Firdaus; Islami, Anggun; Samsuryadi, Samsuryadi
IAES International Journal of Artificial Intelligence (IJ-AI) Vol 13, No 3: September 2024
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/ijai.v13.i3.pp3404-3413

Abstract

Artifacts and unpredictable fetal movements can hinder clear fetal heart imaging during ultrasound scans, complicating anatomical identification. This study presents a new medical imaging approach that combines one-stage instance segmentation with ultrasound (US) video enhancement for precise fetal heart defect detection. This innovation allows real-time identification and timely medical intervention. The study acquired 100 fetal heart US videos from an Indonesian Hospital featuring cardiac septal defects, generating 1,000 frames for training, validation, and testing. Utilizing a combination of the multi-scale input reconstruction network (MIRNet) for image enhancement and YOLOv8l-seg for real-time instance segmentation, the method achieved outstanding validation results, boasting a 99.50% mAP for bounding box prediction and 98.40% for mask prediction. It delivered a remarkable real-time processing speed of 68.4 frames per second. In application to new patients, the method yielded a 65.93% mAP for bounding box prediction and 57.66% for mask prediction. This proposed approach offers a promising solution to early fetal heart defect detection using ultrasound, holding substantial potential for enhancing healthcare outcomes.
TeleOTIVA: Advanced AI-Powered Automated Screening System for Early Detection of Precancerous Lesions Rachmamtullah, Muhammad Naufal; Nurmaini, Siti; Agustiansyah, Patiyus; Sanif, Rizal; Sastradinata, Irawan; Arum, Akhiar Wista; Firdaus, Firdaus; Darmawahyuni, Annisa; Tutuko, Bambang; Sapitri, Ade Iriani; Islami, Anggun
Computer Engineering and Applications Journal Vol 14 No 1 (2025)
Publisher : Universitas Sriwijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.18495/comengapp.v14i1.1197

Abstract

In 2023, the Indonesian Ministry of Health launched the Rencana Aksi Nasional (RAN) to enhance the detection and management of cervical cancer in Indonesia. One of the main pillars in this movement is the implementation of early screening for precancerous lesions aimed at identifying and treating these lesions before they develop into cervical cancer. This effort includes improving public access to healthcare services, providing education and awareness about the importance of early detection, and utilizing the latest technology in screening procedures. It is hoped that, through these targeted and effective interventions, the incidence of cervical cancer can be significantly reduced. This research aims to facilitate the early detection screening process for cervical precancerous lesions, particularly in difficult areas for medical experts to reach. This study also seeks to assist obstetricians and gynecologists in detecting precancerous lesions automatically, quickly, and accurately. By developing an advanced technology-based screening system, it is hoped that early detection of precancerous lesions can be carried out more efficiently, thereby increasing the chances of timely treatment and reducing the incidence of cervical cancer across various regions in Indonesia. This system is designed to provide reliable and user-friendly diagnostic support as it is developed on a mobile platform that can be accessed anytime and anywhere. This research developed a system for early screening called Tele-OTIVA. The Tele-OTIVA application system is an advanced platform that uses artificial intelligence (AI) based approaches to provide optimal services in early detection of precancerous lesions. This application is designed for mobile, allowing users to access and use its advanced features anytime and anywhere. With the integration of AI technology, Tele-OTIVA can detect and analyze cervical precancerous lesions accurately and quickly to provide accurate and efficient screening results. The Tele-OTIVA application system is capable of providing satisfactory detection results. The performance of the proposed model achieves accuracy, sensitivity, and specificity levels above 90%. With this high performance, Tele-OTIVA ensures that the detection of precancerous lesions is carried out with high reliability and precision, instilling greater confidence in healthcare professionals and users during the screening and diagnosis process. The implementation of our application model offers numerous advantages over traditional methods. It significantly enhances efficiency by automating processes, reduces human error through rigorous error-checking mechanisms, and accelerates the processing of large datasets. These improvements streamline operations and ensure more reliable and rapid data analysis.
Identification of Indonesian Authors Using Deep Neural Networks Firdaus; Fahreza, Irvan; Nurmaini, Siti; Darmawahyuni, Annisa; Sapitri, Ade Iriani; Rachmatullah, Muhammad Naufal; Lestari, Suci Dwi; Fachrurrozi, Muhammad; Afrina, Mira; Putra, Bayu Wijaya
Computer Engineering and Applications Journal (ComEngApp) Vol. 11 No. 1 (2022)
Publisher : Universitas Sriwijaya

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

Abstract

Author Name Disambiguation (AND) is a problem that occurs when a set of publications contains ambiguous names of authors, i.e. the same author may appear with different names (synonyms) in other published papers, or author (authors) who may be different who may have the same name (homonym). In this final project, we will design a model with a Deep Neural Network (DNN) classifier. The dataset used in this final project uses primary data sourced from the Scopus website. This research focuses on integrating data from Indonesian authors. Parameters accuracy, sensitivity and precision are standard benchmarks to determine the performance of the method used to solve AND problems. The best DNN classification model achieves 99.9936% Accuracy, 93.1433% Sensitivity, 94.3733% Precision. Then for the highest performance measurement, the case of Non Synonym-Homonym (SH) has 99.9967% Accuracy, 96.7388% Sensitivity, and 97.5102% Precision.
Dermatitis Atopic and Psoriasis Skin Disease Classification by using Convolutional Neural Network Sari, Dwi Mei Rita; Nurmaini, Siti; Rini, Dian Palupi; Sapitri, Ade Iriani
Computer Engineering and Applications Journal (ComEngApp) Vol. 12 No. 1 (2023)
Publisher : Universitas Sriwijaya

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

Abstract

Skin is the one of the body parts that play a large role in human physical body. There are so many functions of the skin such as offering protection against fungal infection, bacteria, allergy, viruses and controls the temperature of the body. But, the reported shown that the skin disease is the most common disease in humans among all age groups and a significant root of infection. The diagnosis of skin diseases involves several tests. Due to this, the diagnosis process is seen to be intensely laborious, time-consuming and requires an extensive understanding aspecially for the skin disease that have similar symptoms. Two skin diseases that have similar symptoms and most misdiagnosed are atopic dermatitis and psoriasis. Convolutional Neural Network for image processing and classifying have been developed for more accurate classification of skin diseases with different architectures. However, the accuracy in determining skin lesions using CNNs is on the average level. The factors that affect the accuracy result of a CNN is the depth where gradients vanished as the network goes deeper. Another factor is the variance in the training set which means the need of the large size of training set. Hence, in this study we tried 10 CNN architecture to get the best result for classifying dermatitis atopic and psoriasis. These are VGG 16, VGG 19, ResNet 50, ResNet 101, MobileNet, MobileNet V2, DenseNet 121, DenseNet 201, Inception and Xception. Experimental result shown that the inception V3 architecture give the best result with accuracy for data testing 84%, accuracy for unseen data 82% and confusion matrix with True positive obtained is 248, True Negative is 61, False positive is 54 and False Negative 298.
Forecasting Of Intensive Care Unit Patient Heart Rate Using Long Short-Term Memory Firdaus; Fachrurrozi, Muhammad; Nurmaini, Siti; Tutuko, Bambang; Rachmatullah, Muhammad Naufal; Darmawahyuni, Annisa; Sapitri, Ade Iriani; Islami, Anggun; Maharani, Masayu Nadila; Putra, Bayu Wijaya
Computer Engineering and Applications Journal (ComEngApp) Vol. 12 No. 3 (2023)
Publisher : Universitas Sriwijaya

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

Abstract

Cardiac arrest remains a critical concern in Intensive Care Units (ICUs), with alarmingly low survival rates. Early prediction of cardiac arrest is challenging due to the complexity of patient data and the temporal nature of ICU care. To address this challenge, we explore the use of Deep Learning (DL) models, specifically Long ShortTerm Memory (LSTM), Bidirectional LSTM (BiLSTM), and Gated Recurrent Unit (GRU), for forecasting ICU patient heart rates. We utilize a dataset extracted from the MIMIC III database, which poses the typical challenges of irregular time series data and missing values. Our research encompasses a comprehensive methodology, including data preprocessing, model development, and performance evaluation. Data preprocessing involves regularizing and imputing missing values, as well as data normalization. The dataset is partitioned into training, testing, and validation sets to facilitate model training and evaluation. Fine-tuning of hyperparameters is conducted to optimize each DL architecture's performance. Our results reveal that the GRU architecture consistently outperforms LSTM and BiLSTM in predicting heart rates, achieving the lowest RMSE and MAE values. The findings underscore the potential of DL models, particularly GRU, in enhancing the early detection of cardiac events in ICU patients.
TeleOTIVA: Advanced AI-Powered Automated Screening System for Early Detection of Precancerous Lesions Rachmatullah, Muhammad Naufal; Nurmaini, Siti; Agustiansyah, Patiyus; Sastradinata, Irawan; Arum, Akhiar Wista; Firdaus; Darmawahyuni, Annisa; Tutuko, Bambang; Sapitri, Ade Iriani; Islami, Anggun
Computer Engineering and Applications Journal (ComEngApp) Vol. 14 No. 1 (2025)
Publisher : Universitas Sriwijaya

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

Abstract

In 2023, the Indonesian Ministry of Health launched the Rencana Aksi Nasional (RAN) to enhance the detection and management of cervical cancer in Indonesia. One of the main pillars in this movement is the implementation of early screening for precancerous lesions aimed at identifying and treating these lesions before they develop into cervical cancer. This effort includes improving public access to healthcare services, providing education and awareness about the importance of early detection, and utilizing the latest technology in screening procedures. It is hoped that, through these targeted and effective interventions, the incidence of cervical cancer can be significantly reduced. This research aims to facilitate the early detection screening process for cervical precancerous lesions, particularly in difficult areas for medical experts to reach. This study also seeks to assist obstetricians and gynecologists in detecting precancerous lesions automatically, quickly, and accurately. By developing an advanced technology-based screening system, it is hoped that early detection of precancerous lesions can be carried out more efficiently, thereby increasing the chances of timely treatment and reducing the incidence of cervical cancer across various regions in Indonesia. This system is designed to provide reliable and user-friendly diagnostic support as it is developed on a mobile platform that can be accessed anytime and anywhere. This research developed a system for early screening called TeleOTIVA. The TeleOTIVA application system is an advanced platform that uses artificial intelligence (AI) based approaches to provide optimal services in early detection of precancerous lesions. This application is designed for mobile, allowing users to access and use its advanced features anytime and anywhere. With the integration of AI technology, TeleOTIVA can detect and analyze cervical precancerous lesions accurately and quickly to provide accurate and efficient screening results. The TeleOTIVA application system is capable of providing satisfactory detection results. The performance of the proposed model achieves accuracy, sensitivity, and specificity levels above 90%. With this high performance, TeleOTIVA ensures that the detection of precancerous lesions is carried out with high reliability and precision, instilling greater confidence in healthcare professionals and users during the screening and diagnosis process. The implementation of our application model offers numerous advantages over traditional methods. It significantly enhances efficiency by automating processes, reduces human error through rigorous error-checking mechanisms, and accelerates the processing of large datasets. These improvements streamline operations and ensure more reliable and rapid data analysis.
Exploration U-Net Architecture for Cervical Precancerous Lesions Segmentation Arum, Akhiar Wista; Rachmatullah, Muhammad Naufal; Tutuko, Bambang; Firdaus; Darmawahyuni, Annisa; Sapitri, Ade Iriani; Islami, Anggun; Ananda, Dea Agustria
Computer Engineering and Applications Journal (ComEngApp) Vol. 14 No. 2 (2025)
Publisher : Universitas Sriwijaya

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

Abstract

The automatic analysis of images for the early detection of cervical cancer relies on the segmentation of cervical precancerous lesions. This paper investigates the incorporation of various CNN-based backbones into a U-Net model for improved segmentation accuracy. A set of twelve backbones was tested, including VGG16, VGG19, ResNet50, ResNext50, EfficientNetB7, InceptionResNetv2, DenseNet201, InceptionV3, MobileNet V2, SE-ResNet50, SE-ResNext50, and SE-Net154. Evaluation metrics were computed using Intersection over Union, pixel accuracy, and Dice coefficient. The findings demonstrate that U-Net with EfficientNetB7 backbone outperforms all other models with an IoU of 73.13%, pixel accuracy of 89.92%, and a Dice coefficient of 77.64%. These results were visually confirmed; segmentation outputs were examined, showing accurate delineation of lesion borders. The dominating performance of EfficientNetB7 was observed to be due to high feature extraction efficiency coupled with powerful spatial information representation. The study is, however, limited by a lack of clinical validation and expert evaluation from trained medical personnel. The results demonstrate the effectiveness of combining the U-Net architecture with advanced CNN backbones towards designing automated systems to analyze medical images.
Deep Learning for ECG-Based Arrhythmia Classification Based on Time-Domain Features Sari, Ririn Purnama; Darmawahyuni, Annisa; Tutuko, Bambang; Firdaus; Rachmatullah, Muhammad Naufal; Sapitri, Ade Iriani; Islami, Anggun; Arum, Akhiar Wista
Computer Engineering and Applications Journal (ComEngApp) Vol. 14 No. 2 (2025)
Publisher : Universitas Sriwijaya

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

Abstract

Arrhythmia is a disturbance in the electrical activity of the heart that can affect the rhythm and duration of the heartbeat. Early detection of arrhythmia is crucial to prevent more serious complications. Electrocardiogram (ECG) is an effective non-invasive diagnostic tool in detecting arrhythmia, but manual detection by experts takes time. To overcome this limitation, this research develops an arrhythmia classification system by utilizing deep learning. This study involves a series of stages, starting from pre-processing, feature extraction, and arrhythmia classification models using convolutional neural networks (CNN) and long short-term memory (LSTM). The results showed that feature extraction successfully improved model efficiency and accuracy. Evaluation of model performance using accuracy, recall, precision, specificity, and F1-score metrics showed that the LSTM model achieved 95% accuracy, 96% recall, 96% precision, 99% specificity, and 96% F1-score, outperforming the CNN model which achieved 91% accuracy, 90% recall, 89% precision, 98% specificity, and 89% F1-score. Thus, these results indicate that the LSTM model is superior in arrhythmia classification.
Residual pixel-wise semantic segmentation for assessing enlarged fetal heart: a preliminary study Roseno, Muhammad Taufik; Nurmaini, Siti; Rini, Dian Palupi; Saputra, Tommy; Mirani, Putri; Rachmatullah, Muhammad Naufal; Darmawahyuni, Annisa; Sapitri, Ade Iriani; Syaputra, Hadi
Bulletin of Electrical Engineering and Informatics Vol 14, No 4: August 2025
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/eei.v14i4.9244

Abstract

The four-chamber view is a crucial scan plane routinely employed in both second-trimester perinatal screening and fetal echocardiographic examinations. Sonographers typically measure biometrics in this plane, such as the cardiothoracic ratio (CTR) and heart axis, to diagnose fetal heart anomalies. However, due to the echocardiographic artifacts, the assessment not only suffers from low efficiency but also inconsistent results depending on the operators’ skills. This study proposes a residual pixel-wise semantic segmentation, which segmented the fetal heart and thoracic contours in a 4-chamber view for assessing an enlarged fetal heart condition. The accuracy of intersection-over-union (IoU) and dice coefficient similarity (DCS) is used for model validation to further regulate the evaluation procedure. We use 1174 US images, comprising about 560 enlarged heart images, and about 614 normal heart images. Out of these data, 248 images are used for unseen data, and the remaining for training/validation processes. The performance of the proposed model, when tested on unseen data, achieved satisfactory results with 97.71% accuracy, 90.36% IoU, and 94.93% DCS. These metrics collectively demonstrate the satisfactory performance of the proposed model compared to existing segmentation models. The outcomes underscore that the proposed model establishes a state-of-the-art standard for enlarged fetal heart detection.
Enhancing ultrasound image quality using deep structure of residual network Sapitri, Ade Iriani; Nurmaini, Siti; Rachmatullah, Muhammad Naufal; Darmawahyuni, Annisa; Firdaus, Firdaus; Islami, Anggun; Tutuko, Bambang; Arum, Akhiar Wista
International Journal of Electrical and Computer Engineering (IJECE) Vol 15, No 4: August 2025
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/ijece.v15i4.pp3779-3794

Abstract

Ultrasonography, a medical imaging technique, is often affected by various types of noise and low brightness, which can result in low image quality. These drawbacks can significantly impede accurate interpretation and hinder effective medical diagnoses. Therefore, improving image quality is an essential aspect of the field of ultrasound systems. This study aims to enhance the quality of ultrasound images using deep learning (DL). The experiment is conducted using a custom dataset consisting of 2,175 infant heart ultrasound images collected from Indonesian hospitals, and the model is subsequently generalized using other datasets. We propose enhanced deep residual network combined convolutional neural networks (EDR-CNNs) to improve the image quality. After the enhancement process, our model achieved peak signal-to-noise ratio (PSNR) and structural similarity index metrics (SSIM) scores of 38.35 and 0.92 respectively, outperforming other methods. The benchmarking with other ultrasound medical images indicates that our proposed model produces good performance, as evidenced by higher PSNR, lower SSIM, a decrease in mean square error (MSE), and a lower contrast improvement index (CII). In conclusion, this study encapsulates the forthcoming trends in advancing low-illumination image enhancement, along with exploring the prevailing challenges and potential directions for further research.