Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal Jurnal Infra
Abraham Imanuel
Program Studi Teknik Informatika, Universitas Kristen Petra Surabaya
Computer Science & IT

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

Found 1 Documents
Search

 1 
Penerapan Convolutional Neural Network dengan Pre-Trained Model Xception untuk Meningkatkan Akurasi dalam Mengidentifikasi Jenis Ras Kucing Abraham Imanuel; Djoni Haryadi Setiabudi
Jurnal Infra Vol 10, No 2 (2022)
Publisher : Universitas Kristen Petra

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Raising pets is a common thing that humans often do. Cats are one of the domestic pets that are fancied by humans. But, raising a cat is not an easy task. This is because every cat’s breed has its own characteristics that will affect its type of raising. Because of that, there is a need for a system that can identify a cat’s breed to help someone in deciding which type of cat is suitable for him/her. In the past, there was also research about cat’s breed detection using SSD Mobilenet_v1 FPN method, but the accuracy was not high enough, which was 81.74%. This thesis will be done with the implementation of transfer learning method on Pre-Trained Convolutional Neural Network Xception, which is a CNN Model that is inspired by CNN Model Inception. CNN Model Xception is an Inception Model that replaces the use of Inception modules with depth wise separable convolutions. CNN Model Xception is used in this thesis with the purpose of increasing the accuracy of cat’s breed detection. Output of the system shows that the highest accuracy that could be made in detecting cat’s breeds on The Oxford-IIIT Pet Dataset is 89.58% or 0.8958. Compared to SSD Mobilenet_v1 FPN method, which accuracy was 81.74%, implementing Xception method gives an increase in the accuracy for 7.84%. Other than that, it is also found that the dataset quality has an impact on model’s accuracy.
Co-Authors Djoni Haryadi Setiabudi