Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal Smatika Jurnal : STIKI Informatika Jurnal
Adryan Putra Pratama
UIN Sunan Gunung Djati Bandung, Indonesia
Computer Science & IT

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Klasifikasi Ras Kucing Dengan Pendekatan Convolutional Neural Networks Menggunakan Arsitektur Inception V4 Adryan Putra Pratama; Jumadi Jumadi; Eva Nurlatifah
SMATIKA JURNAL : STIKI Informatika Jurnal Vol 15 No 02 (2025): SMATIKA Jurnal : STIKI Informatika Jurnal
Publisher : LPPM UBHINUS MALANG

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32664/smatika.v15i02.1547

Abstract

Classifying cat breeds based on images presents challenges due to subtle differences in appearance among breeds and environmental influences. This study developed an automated classification system utilizing the Inception V4 architecture with a CRISP-DM approach, encompassing business understanding, data preparation, modeling, evaluation, and deployment. The dataset used was derived from the Oxford IIIT Pet Dataset, covering 12 popular cat breeds, and underwent cleaning, augmentation, normalization, and partitioning into training (80%) and validation (20%) datasets. The model was trained over 25 epochs, achieving a highest validation accuracy of 93.31% with average precision, recall, and f1-score of 93%. The system was implemented as a Flask-based web application, enabling real-time classification through image uploads. While overall performance was strong, certain breeds such as Bengal exhibited potential for further improvement.  The findings demonstrate the model's significant potential to support pet health diagnosis and breed conservation efforts. This study contributes substantially to the development of image-based classification technology, with recommendations for performance enhancements through GAN-based data augmentation and testing on more diverse datasets to improve generalizability.
Co-Authors Eva Nurlatifah Jumadi Jumadi