Skripsi

PENGENALAN PRODUK E-COMMERCE BERBASIS GAMBAR SECARA OTOMATIS MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK (CNN)

---

E-commerce has become a key pillar of the digital economy, supported by broad internet access and modern technology. However, text-based searches often cause difficulties, especially for products with complex descriptions. This study develops an automatic image-based e-commerce product recognition system using Convolutional Neural Network (CNN) with the VGG16 architecture. The Fashion Product Images (Small) dataset from Kaggle, containing 44,441 images, was filtered to 19,520 valid images, then augmented and balanced to 25,000 images evenly distributed across 10 categories. The dataset was split into 70% training, 15% validation, and 15% testing. Four VGG16 configurations were tested with different hyperparameters. Model 4 (learning rate 0.00005, dropout 0.3, batch size 32, rotation range 20°, L2 0.01) achieved the highest accuracy of 99.04%, with macro average precision, recall, and F1-score of 0.99. The system was implemented in a Streamlit-based web application that allows users to upload product images and obtain real-time recognition results.

---

Availability

Detail Information

Series Title

-

Call Number

T1919602026

Publisher

Indralaya : Prodi Teknik Informatika, Fakultas Ilmu Komputer Universitas Sriwijaya., 2026

Collation

x, 111 hlm.; ilus.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

006.370 7

Content Type

Text

Media Type

unmediated

Carrier Type

-

Edition

-

Subject(s)

Prodi Teknik Informatika
Convolutional Neural Network

Specific Detail Info

-

Statement of Responsibility

KA

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment