Claim Missing Document
Check
Articles

Found 32 Documents
Search

Sistem Informasi Deteksi Deepfake Video Promosi Affiliate Menggunakan Arsitektur Inception-ResNet v2 Joko Dwi Mulyanto; Supriatiningsih Supriatiningsih; Ubaidillah Ubaidillah
Informatics and Computer Engineering Journal Vol 6 No 2 (2026): Periode Juli 2026
Publisher : Lembaga Penelitian dan Pengabdian Masyarakat (LPPM) Universitas Bina Sarana Informatika

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31294/icej.v6i2.13694

Abstract

Maraknya pemanfaatan video pendek dalam platform affiliate marketing saat ini menghadapi ancaman siber baru berupa teknologi deepfake. Manipulasi wajah tokoh publik atau influencer oleh pihak tidak bertanggung jawab untuk mengejar komisi afiliasi berpotensi merusak integritas sistem, merugikan konsumen, dan menurunkan reputasi platform e-commerce. Penelitian ini bertujuan untuk merancang sebuah subsistem tata kelola konten (Content Governance IS) otomatis guna mendeteksi dan memitigasi penyebaran video deepfake promosi produk. Metode yang diusulkan mengintegrasikan algoritma Multi-task Cascaded Convolutional Networks (MTCNN) pada tahap preprocessing untuk mengekstrak Region of Interest (ROI) wajah secara dinamis ke dalam matriks 160 X 160 piksel melalui lingkungan Google Colab. Selanjutnya, klasifikasi biner dilakukan memanfaatkan teknik Transfer Learning berbasis arsitektur Deep Learning Inception-ResNet v2 yang dikombinasikan dengan Global Average Pooling serta lapisan dropout (rate=0.5) untuk mencegah overfitting. Pengambilan keputusan pada sistem informasi ini menerapkan Three-Tier Decision Framework dengan pembagian zona otomatis (Approved, Pending untuk Audit, dan Rejected). Sesuai hipotesis, implementasi model ini mampu menghasilkan deteksi dengan tingkat akurasi yang tinggi serta efisiensi waktu pemrosesan komputasi yang sangat cepat (waktu inferensi < 1 detik per video). Hasil penelitian ini diharapkan dapat memberikan kontribusi signifikan berupa model arsitektur sistem informasi keamanan konten yang tangguh, adaptif, dan siap diintegrasikan sebagai API pada sistem manajemen konten skala besar.
DIGITALISASI MANAJEMEN USAHA DAN PEMASARAN UNTUK MENINGKATKAN KINERJA UMKM RUSMY ROSE TEH BUNGA TELANG Joko Dwi Mulyanto; Ina Maryani; Suripah Suripah; Nisrina Lukluk'il Numa; Naufal Haidar Azhar
Jurnal Abdi Insani Vol 13 No 3 (2026): Jurnal Abdi Insani
Publisher : Universitas Mataram

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29303/abdiinsani.v13i3.3437

Abstract

Rusmy Rose Teh Bunga Telang UMKM has potential in herbal products but faces serious obstacles in management and marketing. The business management system, particularly financial and stock recording, is still done manually and is unstructured. In addition, marketing activities are very limited to offline sales and do not utilize digital platforms, resulting in stagnant turnover and a narrow market reach. This community service activity aims to digitize business management and marketing to improve the performance and competitiveness of MSMEs. The method used is participatory empowerment. The activity was carried out in three main stages: assessment (FGD), training (digital management, branding, e-commerce, AI), and technology implementation assistance. The results of the activity showed significant changes in the partners. In terms of management, partners have implemented digital financial and stock records using Google Sheets. In terms of marketing, partners now have and manage e-commerce accounts (Shopee, Tokopedia) and websites. Partners are also able to produce independent promotional content using Canva and AI. The main impact is an increase in average monthly sales turnover of 20-30%. Additionally, the market reach has been successfully expanded to new areas such as Purbalingga, Banjarnegara, and Cilacap. It is concluded that comprehensive digitalization of management and marketing has proven effective in improving the performance and competitiveness of Rusmy Rose UMKM