No image available for this title

Tugas Akhir Informatika

KLASIFIKASI TINGKAT KESEGARAN CUMI – CUMI BERDASARKAN FITUR WARNA DAN TEKSTUR DENGAN MENGGUNAKAN METODE SUPPORT VECTOR MACHINE

Marine animals are very susceptible to decay. The traditional method that is often used to distinguish the freshness of
squid by local people is from the body color and smell of the squid. This method is very simple but has many shortcomings
in distinguishing freshness from squid. The drawback of this method lies in the understanding and level of accuracy of
each person who is different. So it is necessary to create a system that can distinguish the freshness level of squid
automatically only from the image of the squid. In this study, a system model was developed that can classify the
freshness level of squid using the Support Vector Mahine (SVM) method. The GLCM and histogram methods as well as
the HSI color space are used for texture and color feature extraction. This study uses three types of classification. The
total data used in this study are 360 body images of squid that have been cropped and resized by 128 x 128 pixels for
the treatment type class. The total data for the freshness class with three types of classes is 495 and the total data for
the freshness class with two types of classes is 330. In this study, the process of cropping, augmentation, resizing and
conversion of color space in the dataset was carried out. The distribution of training data and test data is 70:30. The
highest accuracy obtained is 67.75%

Ketersediaan

Tidak ada salinan data

Informasi Detail
Judul Seri
-
No. Panggil
006.31.Fat;ki
Penerbit
Universitas Mataram : Fakultas Teknik Unram.,
Deskripsi Fisik
xii,58 Hlm; Illus;21x29 cm
Bahasa
Indonesia
ISBN/ISSN
-
Klasifikasi
006.31
Tipe Isi
other
Tipe Media
computer
Tipe Pembawa
audio disc
Edisi
Edisi 1 Jilid 1
Subjek
Kecerdasan buatan
Squids Freshness, SVM, GLCM, HSI
Multi-Class
Info Detail Spesifik
-
Pernyataan Tanggungjawab
Versi lain/terkait

Tidak tersedia versi lain

Lampiran Berkas
Komentar
Anda harus login sebelum memberikan komentar