Article Per Year (5 Year)
p-Index From 2021 - 2026
0.444
P-Index
This Author published in this journals
All Journal Infotech Journal Journal Medical Informatics Technology
Wildan Zhilal Manafi
Universitas Majalengka
Computer Science & IT Dentistry Engineering Medicine & Pharmacology Public Health

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

Found 2 Documents
Search

 1 
EXPLAINABLE DEEP LEARNING FOR BEEF FRESHNESS CLASSIFICATION USING GRAD-CAM VISUALIZATION Ade Bastian; Ardi Mardiana; Billy Adrian Fernanda; Harun Sujadi; Abrar Wahid; Riri Nurazizah; Wildan Zhilal Manafi
INFOTECH journal Vol. 11 No. 2 (2025)
Publisher : Universitas Majalengka

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31949/infotech.v11i2.16897

Abstract

Kesegaran daging sapi merupakan faktor kritis bagi keamanan pangan di Indonesia, mengingat tingginya tingkat konsumsi dan impor komoditas ini. Metode penilaian kesegaran tradisional seringkali lambat, destruktif (merusak), atau bias secara subjektif. Penelitian ini bertujuan untuk mengembangkan model Deep Learning yang tidak hanya akurat dalam mengklasifikasikan kesegaran daging sapi (Segar, Setengah Segar, Busuk) tetapi juga dapat dijelaskan (explainable) dalam proses pengambilan keputusannya. Kami menerapkan Transfer Learning menggunakan arsitektur Convolutional Neural Network (CNN) yang ringan, yaitu MobileNetV2, pada dataset yang terdiri dari 2.266 citra daging yang telah diaugmentasi. Untuk mengatasi sifat "black-box" dari CNN, Gradient-weighted Class Activation Mapping (Grad-CAM) diimplementasikan untuk memvisualisasikan area fokus model. Hasil eksperimen menunjukkan bahwa model kami yang telah di-fine-tune mencapai akurasi validasi yang tinggi (96,01%), dengan presisi sempurna (100%) untuk kelas 'Busuk' (Spoiled), memastikan tidak ada daging busuk yang salah diklasifikasikan sebagai daging segar. Analisis Grad-CAM lebih lanjut memvalidasi bahwa model mendasarkan keputusannya pada fitur visual yang relevan secara biologis, seperti pola perubahan warna dan tekstur permukaan, bukan pada noise latar belakang. Temuan ini mengonfirmasi potensi integrasi CNN ringan dengan XAI untuk sistem kontrol kualitas yang andal, non-destruktif, dan transparan dalam industri pangan.
Mobile Web App Development for Diabetic Foot Screening Using Inlow’s 60-Second Screen with Automated Risk Classification Suhendri; Wildan Zhilal Manafi; Bayu Reviyadi; Sri Rahayu; Iin Karmila Septiani; Mita Nurmala
Journal Medical Informatics Technology Volume 4 No. 2, June 2026
Publisher : SAFE-Network

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37034/medinftech.v4i2.152

Abstract

Diabetic foot complications constitute a major contributor to preventable lower-extremity amputation, yet primary care screening remains inconsistent due to the absence of integrated digital tools implementing validated clinical protocols. This study presents the design, implementation, and system-centric evaluation of Podiatrix, a mobile web application that operationalizes Inlow's 60-Second Diabetic Foot Screen through an automated, condition-based clinical workflow. Unlike existing tools that address isolated screening criteria, Podiatrix implements all seven Inlow criteria within a unified five-step wizard and applies a deterministic hierarchical classification engine that directly mirrors the original Inlow protocol logic rather than relying on fixed score thresholds. The system was evaluated using three complementary methods: black-box testing across 50 simulated clinical scenarios, Nielsen's heuristic usability evaluation conducted by three independent evaluators, and performance load testing using Apache JMeter under concurrent user conditions. Results demonstrated 100% classification accuracy (50/50 scenarios) matching manual Inlow protocol interpretation, an average heuristic severity score of 1.15 out of 4 indicating high usability, and a mean response time of 820 ms with less than 1% error rate under 100 concurrent users. These findings confirm that Podiatrix provides a computationally robust, highly usable, and scalable digital infrastructure that lays the groundwork for future prospective clinical trials in primary care and community health settings.
Co-Authors Abrar Wahid Ade Bastian Ardi Mardiana Bayu Reviyadi Billy Adrian Fernanda Harun Sujadi Iin Karmila Septiani Mita Nurmala Riri Nurazizah Sri Rahayu Suhendri Suhendri