Sistem Klasifikasi Citra Makanan Menggunakan Representasi Anti Textons dan K-Nearest Neighbour

Septiana, Nuning (2017) Sistem Klasifikasi Citra Makanan Menggunakan Representasi Anti Textons dan K-Nearest Neighbour. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 5113100143-Undergraduate-Theses.pdf]
Preview
Text
5113100143-Undergraduate-Theses.pdf - Published Version

Download (3MB) | Preview

Abstract

Saat ini, masyarakat mudah mendapatkan citra. Citra yang didapat bisa diambil sendiri lewat media seperti handphone atau dari internet. Kemudahan dalam mendapatkan citra menimbulkan kebutuhan untuk mengetahui informasi citra berdasarkan kontennya. Salah satu kebutuhannya yaitu klasifikasi citra. Klasifikasi citra bisa diterapkan pada diatery assesment, evaluasi komprehensif tentang makanan yang dikonsumsi oleh manusia.
Pada tugas akhir ini dibangun sistem klasifikasi citra menggunakan fitur anti-texton dan klasifikasi K-NN. Anti-Texton dihitung dengan cara mengukur jarak spasial dari texton. Texton sendiri berasal dari Texton Library yang didapatkan dengan menggunakan K-Means clustering pada data training. Input dari sistem ini adalah citra makanan dan output-nya adalah kelas makanan.
Berdasarkan uji coba, akurasi tertinggi mencapai 77,4%. Hasil ini dipengaruhi oleh kondisi dataset dimana terdapat beberapa data citra yang punya kemiripan meskipun kelasnya berbeda.

============================================================
In this era, it is easy to get an image. People can get it by media like phone or from internet. The easiness in getting image make people become need about the information of image based on content. One of the exampe is image classification. Image classification can be implemented in Dietary assessment, a comprehensive evaluation of a person's food intake.
In this final project the system is being constructed using the Anti-Texton representation and K-NN. Anti-Texton is a value of spatial distance between texton. The texton itself is gained from Texton Library which is obtained by K-Means clustering of training data. The system need food image as the input and will decide the class that suit to the food image as the output
Based on the test results, the maximum average accuracy of this classification is 60.2%. This result is caused by the condition of the dataset which have some similar image eventhough it is different class.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Anti Texton, K-Means clustering, Texton library, K-NN, citra makanan, Dietary assesment, food images
Subjects: Z Bibliography. Library Science. Information Resources > ZA Information resources > Z699.5 Information storage and retrieval systems
Divisions: Faculty of Information Technology > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User: - Nuning Septiana
Date Deposited: 08 Nov 2017 07:12
Last Modified: 05 Mar 2019 07:29
URI: http://repository.its.ac.id/id/eprint/42521

Actions (login required)

View Item View Item