ACADSTAFF UGM

CREATION
Title : Penggunaan Pre-trained Model untuk Klasifikasi Kualitas Sekrup
Author :

MUHAMMAD RAFIF (1) Diyah Utami Kusumaning Putri, S.Kom., M.Sc., M.Cs. (2) Lukman Awaludin, S.Si., M.Cs. (3)
Date : 0 2022
Keyword : Inspeksi kualitas produk,Pre-trained model,Augmentasi citra Inspeksi kualitas produk,Pre-trained model,Augmentasi citra
Abstract : Image-based product quality inspection is important for the manufacturing industry. Most of these tasks are still carried out by humans who have low units per hour. Conventional methods for image inspection still rely on feature-based methods, which have difficulties in generalizing and extracting features. This problem is solved by usingCNN method, but CNN requires large data andlong training time. Pre-trained modeland image augmentation can solve the problems in the previous methods. However, there has been no research that has completely examined and compared the performance of various pre-trained modelsand variations of image augmentation for image inspection task ofmanufactured productsquality.The research process uses a multi-class and binary-class screw dataset on 33 pre-trained modeltypesand 8 image augmentationtypes. Testing the pre-trained model using a combined dataset of allimage augmentationtypes. Modelwith the highest accuracy isEfficientNetV2-L for multi class dataset(97.8%) and VGG-19 for binary class dataset(96.5%). Image augmentation with the highest significance to model performance is blur, with an accuracy of 81.1% in multi class and 92% in binary class. The whole process of testing pre-trained modeland image augmentation went well.
Group of Knowledge : Ilmu Komputer
Original Language : Bahasa Indonesia
Level : Nasional
Status :
Published
Document
No Title Document Type Action
1 Penggunaan Pre-trained Model untuk Klasifikasi Kualitas Sekrup.pdf
Document Type : [PAK] Full Dokumen
[PAK] Full Dokumen View
2 Bukti Akreditasi.pdf
Document Type : Dokumen Pendukung Karya Ilmiah (Hibah, Publikasi, Penelitian, Pengabdian)
Dokumen Pendukung Karya Ilmiah (Hibah, Publikasi, Penelitian, Pengabdian) View
3 Bukti Korespondensi IJEIS.pdf
Document Type : [PAK] Bukti Korespondensi Penulis
[PAK] Bukti Korespondensi Penulis View