Article Per Year (5 Year)
p-Index From 2021 - 2026
0.408
P-Index
This Author published in this journals
All Journal Jurnal Penelitian Sistem Informasi Journal of Embedded Systems, Security and Intelligent Systems
Tri Kustanti Rahayu
Universitas Musamus
Computer Science & IT Decision Sciences, Operations Research & Management

Published : 2 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 2 Documents
Search

 1 
Accuracy–Efficiency Trade-off Analysis of Five Lightweight CNN Architectures for Mobile-Deployable Corn Leaf Disease Classification Jarot Budiasto; Hasanudin Jayawardana; Tri Kustanti Rahayu; Tatik Melinda Tallulembang
Journal of Embedded Systems, Security and Intelligent Systems Vol 7 No 3 (2026): September 2026
Publisher : Program Studi Teknik Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59562/jessi.v7i3.12264

Abstract

Purpose – Corn leaf disease diagnosis in resource-constrained agricultural settings requires mobile-deployable models that maintain a practical balance between classification accuracy, model size, and on-device latency. This study aims to provide empirical guidance for selecting lightweight Convolutional Neural Network (CNN) architectures by systematically analyzing the accuracy–efficiency trade-offs of five models for corn leaf disease classification. Design/methods/approach – MobileNetV2, MobileNetV3-Small, MobileNetV3-Large, EfficientNetB0, and NASNetMobile were evaluated on the PlantVillage Corn dataset comprising 4,188 images across four classes under identical experimental settings. The models were trained using a two-phase strategy and converted into standard and dynamic-range quantized TensorFlow Lite formats. Evaluation covered classification accuracy, macro F1-score, model size, Android on-device inference latency, Pareto frontier and radar analyses, and pairwise McNemar's tests with Yates continuity correction. Findings – EfficientNetB0 achieved the highest accuracy (95.25%) and macro F1-score (93.77%). MobileNetV3-Small offered the strongest efficiency under the tested Android CPU setting, reaching 94.54% accuracy with a 1.18 MB dynamic-range quantized TensorFlow Lite model and 3.89 ± 0.04 ms standard inference. The top three models were statistically comparable (p = 0.6625-1.0000). Research implications/limitations – Standard TensorFlow Lite is preferable for low-latency Android CPU deployment, whereas dynamic-range quantized TensorFlow Lite supports storage-constrained offline distribution. However, the findings are limited to the PlantVillage benchmark dataset and testing on a single mid-range Android device. Originality/value – This study integrates lightweight CNN benchmarking, TensorFlow Lite deployment, real-device Android testing, accuracy–efficiency trade-off analysis, and statistical validation to support evidence-based mobile agricultural AI model selection.
Comparative Analysis of Data Augmentation Strategies for Corn Leaf Disease Classification Using Transfer Learning Muhammad Hasbi; Jarot Budiasto; Hasanudin Jayawardana; Tri Kustanti Rahayu
JURNAL PENELITIAN SISTEM INFORMASI (JPSI) Vol. 4 No. 4 (2026): November : JURNAL PENELITIAN SISTEM INFORMASI
Publisher : Institut Teknologi dan Bisnis (ITB) Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.54066/jpsi.v4i4.5091

Abstract

The choice of data augmentation can change how a transfer-learning model represents disease symptoms, particularly when classes differ through subtle lesion color and shape. This study evaluated five training conditions for MobileNetV2-based corn leaf disease classification: no augmentation, geometric augmentation, color augmentation, combined augmentation, and strong augmentation. Experiments used 4,188 PlantVillage corn leaf images from the Blight, Common Rust, Gray Leaf Spot, and Healthy classes. A fixed stratified 70:20:10 split and the same two-stage training protocol were applied in every condition, so the augmentation policy was the only experimental factor. Performance was assessed from test accuracy, precision, recall, F1-score, confusion matrices, and error counts. Color augmentation produced the strongest result, reaching 92.60% accuracy, a 91.42% macro F1-score, a 92.52% weighted F1-score, and 31 misclassified images. It also raised the Gray Leaf Spot F1-score from 82.35% without augmentation to 84.91%. Geometric augmentation offered no gain, whereas the combined and strong policies reduced overall performance. For this dataset, moderate brightness and contrast variation preserved disease structure more effectively than broader spatial or aggressive transformations.
Co-Authors Hasanudin Jayawardana JAROT BUDIASTO Muhammad Hasbi Tatik Melinda Tallulembang