Claim Missing Document
Check
Articles

Found 4 Documents
Search

Segmentasi Semantik pada Citra Hama Leafblast Menggunakan Unet dan Optimasi Hyperband Moch. Nasheh Annafii; Oddy Virgantara Putra; Triana Harmini; Niken Trisnaningrum
Prosiding Seminar Sains Nasional dan Teknologi Vol 12, No 1 (2022): VOL 12, NO 1 (2022): PROSIDING SEMINAR NASIONAL SAINS DAN TEKNOLOGI
Publisher : Fakultas Teknik Universitas Wahid Hasyim

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36499/psnst.v12i1.7230

Abstract

Padi menjadi konsumsi primer di Indonesia. Penyakit padi menjadi salah satu faktor yang menyebabkan menurunnya jumlah produksi padi. Meningkatnya konsumsi beras menjadi masalah dengan menurunnya jumlah produksi padi pada tahun 2021. Luasnya lahan dan lambatnya proses identifikasi keparahan menjadikan kurang maksimalnya penanganan penyakit padi, yang berujung tidak maksimalnya hasil panen bahkan terancam gagal panen. Penelitian ini berupaya untuk mensegmen daun padi yang terkena hama leafblast dengan model yang dioptimalkan. Penelitian ini menggunakan metode algoritma Convolutional Neural Network dengan model UNet yang ditingkatkan dengan optimasi model Hyperband optimization. Dengan banyaknya penelitian mengenai UNet, UNet menjadi populer dan berkembang dengan pesat. Perkembangan yang pesat ini ditandai dengan banyaknya penelitian yang menggunakan UNet dan banyaknya modifikasi yang terus dikembangkan. Dataset yang digunakan pada penelitian ini merupakan murni hasil observasi peneliti dan telah divalidasi oleh ahli, dengan total 300 data asli dan data label. Dalam model yang digunakan, digambarkan terdapat bagian encoder dan decoder yang masing masing memiliki beberapa blok konvolusi. Hasil yang diperoleh dari model yang sudah dioptimasi terbukti 3 kali lebih ringan dengan perbandingan jumlah parameter yang cukup signifikan dan hasil valuasi akurasi mencapai 97.72%.
Klasifikasi Tingkat Keparahan Penyakit Leafblast Tanaman Padi Menggunakan MobileNetv2 Imam Fauzi Annur; Jumhurul Umami; Moch. Nasheh Annafii; Niken Trisnaningrum; Oddy Virgantara Putra
Fountain of Informatics Journal Vol. 8 No. 1 (2023): Mei
Publisher : Universitas Darussalam Gontor

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21111/fij.v8i1.9419

Abstract

AbstrakPadi merupakan tanaman pangan pokok di Indonesia, dan produksinya merupakan kunci ketahanan pangan negara. Keberhasilan panen merupakan faktor penting dalam pencegahan impor bahan pangan pokok. Tantangan terbesar dalam memanen tanaman adalah adanya virus, jamur, dan hama yang dapat merusak tanaman. Penelitian ini bertujuan untuk membuat sistem klasifikasi tingkat keparahan penyakit daun pada tanaman padi yang terkena penyakit blas daun dengan bantuan algoritma machine learning. MobileNetV2 adalah arsitektur Convolutional Neural Network (CNN) yang menggunakan Depthwise Separable Convolution untuk membangun model yang ringan dan dirancang untuk mengatasi proses yang memiliki resource yang berlebih. Dataset yang digunakan pada penelitian ini merupakan hasil murni observasi peneliti yang sudah divalidasi oleh ahli dengan total 300 data asli. Model MobileNetV2 ternyata sangat berhasil dalam mengklasifikasikan objek, dengan akurasi 78,33%. dengan hasil penelitian ini, petani dapat terbantu dalam mengenali tingkat keparahan penyakit leafblast pada tanaman padi sehingga pemberian bahan kimia berupa fungisida sesuai dengan dosis anjuran tingkat keparahan. Kata kunci: Klasifikasi, leafblast, padi, citra, model pre-trained, MobileNetV2. Abstract[Classification Of Rice Blast Disease Using MobileNetV2] Rice is a staple food crop in Indonesia, and its production is key to the country's food security. Successful harvesting is an important factor in preventing imports of staple foods. The biggest challenge in harvesting crops is the presence of viruses, fungi, and pests that can damage plants. This research aims to create a classification system for leaf disease severity in rice plants affected by leaf blast disease with the help of machine learning algorithms. MobileNetV2 is a Convolutional Neural Network (CNN) architecture that uses Depthwise Separable Convolution to build lightweight models and is designed to overcome processes that have excessive resources. The dataset used in this study is the result of pure researcher observations that have been validated by experts with a total of 300 original data. The MobileNetV2 model turned out to be very successful in classifying objects, with an accuracy of 78.33%. with the results of this study, farmers can be helped in recognizing the severity of leafblast disease in rice plants so that the provision of chemicals in the form of fungicides in accordance with the recommended dose of severity.Keywords: Classification, leafblast, rice, image, pre-trained model, MobileNetV2
Human Digital Twin Modeling for Cardiovascular System Herman, Herman; Annafii, Moch. Nasheh; Kunta Biddinika, Muhammad; Fitriah, Fitriah
Scientific Journal of Informatics Vol. 12 No. 1: February 2025
Publisher : Universitas Negeri Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15294/sji.v11i4.16012

Abstract

Purpose: The cardiovascular system is a vital system responsible for the distribution of oxygen and nutrients throughout the body. The complexity of interactions between the heart and blood vessels often presents challenges in monitoring and analyzing health conditions. The research proposes the development of a Human Digital Twin (HDT) for the cardiovascular system through application of two different modelling approaches geometric modeling and physic-based modeling. Through this model physical conditions can be represented and real time data integrated to offer insights into the dynamics of the cardiovascular system. Methods: This model development is based on two major components: a geometric modeling and a physic-based modeling. The geometric model is done in 3D to show the structure of the heart in detail, while the physical-based model is tabulated with different measurable physical parameters in the cardiovascular system, such as blood pressure and flow rate. This information is integrated into the Five Dimension Digital Twin model, including physical, virtual, data, connection, and service dimensions for the accurate simulation of cardiovascular conditions. Result: Results confirm that the Five-Dimensional Digital Twin (DT) could give further development to how the dynamics of the cardiovascular system behave, possibly in real-time updates on conditions and a supply of data that is far more detailed in view of analyzing risk and further representation of specific cardiovascular disorders while providing personalized medical support. Novelty: The Five-Dimensional Human Digital Twin Model (HDTM) developed in this research introduces novel innovations in the monitoring and simulation of the cardiovascular system through the application of geometric and physic-based modeling techniques. This approach offers a higher level of detail, compared to previous models, and added value for the advancement of health technology by integrating real time data into the simulations. This model serves not only as an advanced analytical tool but also as a reference for further research on DT technology in the medical field.
HUMAN DIGITAL TWIN MODELING FOR ADVANCING ARRHYTHMIA TREATMENT herman herman; Moch. Nasheh Annafii; Muhammad Kunta Biddinika
JIKO (Jurnal Informatika dan Komputer) Vol 8 No 3 (2025)
Publisher : Program Studi Teknik Informatika Universitas Khairun

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33387/jiko.v8i3.10656

Abstract

Heart disease in all its forms remains a significant health threat. Arrhythmia is a type of heart disease whose diagnosis and treatment still primarily rely on conventional electrocardiogram-based diagnosis. However, this approach is limited, as it is reactive and captures cardiac conditions only at the time of electrocardiogram measurement, making it unable to continuously and individually monitor arrhythmia progression for each patient. This study explores digital twin technology and develops human digital twin models for the treatment of arrhythmia patients. The modeling framework integrates three core components: geometrical modeling, physical modeling, and data-driven modeling to represent the human heart and cardiovascular system in a digital environment. The output of this integrative process has been implemented in the initial prototype of the Human Digital Twin Cockpit, which is designed to treat arrhythmia. This prototype enhances the existing diagnosis and treatment, and also incorporates a proactive simulation system. Evaluation and system testing have successfully demonstrated their ability to integrate geometric data from medical imaging and physical data from electrophysiological sensors to predict arrhythmia in various scenarios