Penilaian Estetika Foto Makanan Berdasarkan Komposisi dan Tekstur Visual

Jauhari, Haniif Ahmad (2024) Penilaian Estetika Foto Makanan Berdasarkan Komposisi dan Tekstur Visual. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5025201224-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (10MB) | Request a copy

Abstract

Dengan meningkatnya popularitas media sosial dan platform berbagi foto seperti Instagram, banyak orang semakin sering mengunggah foto makanan. Foto makanan yang menarik sering kali menjadi viral, menciptakan ekspektasi yang lebih tinggi terhadap estetika makanan. Namun, mengevaluasi foto makanan dapat menjadi suatu tantangan tersendiri. Pada penelitian sebelumnya, telah diuji peran fitur komposisi dan warna dalam menilai estetika foto makanan. Hasil penelitian tesebut menunjukkan bahwa fitur komposisi memiliki pengaruh signifikan dalam menentukan nilai estetika makanan, sementara fitur warna kurang berpengaruh, terutama ketika digabungkan dengan fitur komposisi. Oleh karena itu, penelitian ini menguji peran fitur tekstur dalam menilai estetika foto makanan dan bagaimana pengaruhnya ketika digabungkan dengan fitur komposisi yang telah terbukti berpengaruh dalam menilai estetika foto makanan. Penilitian ini menggunakan metode ekstraksi fitur komposisi menggunakan modul SAMP-Net, serta metode ekstraksi fitur tekstur menggunakan Gray Level Co-occurrence Matrix (GLCM) dan transformasi wavelet. Eksperimen dilakukan pada dataset publik Gourmet Photography Dataset yang terdiri dari 24.000 foto makanan. Model klasifikasi yang digunakan adalah Support Vector Machine (SVM), Random Forest, Multilayer Perceptron (MLP), dan VGG16+MLP. Hasil penelitian dievaluasi menggunakan metrik seperti akurasi, precision, recall, dan f1-score untuk setiap model klasifikasi yang digunakan. Dari hasil ablasi fitur, ditemukan bahwa fitur tekstur GLCM memiliki pengaruh positif paling besar terhadap akurasi model klasifikasi. Di antara keempat model klasifikasi yang digunakan, model VGG16+MLP memiliki kinerja terbaik dengan akurasi sebesar 0,9108, precision sebesar 0,9072, recall sebesar 0,9321, dan f1-score sebesar 0,9195.
============================================================
With the increasing popularity of social media and photo-sharing platforms such as Instagram, people are increasingly uploading food photos. Attractive food photos often go viral, creating higher expectations for food aesthetics. However, evaluating food photos can be challenging. In a previous study, the role of composition and color features in assessing the aesthetics of food photos was examined. The results show that composition features have a significant influence in determining the aesthetic value of food, while color features have less influence, especially when combined with composition features. Therefore, this study examines the role of texture features in assessing the aesthetics of food photos and how they affect them when combined with compositional features that have been shown to be influential in assessing the aesthetics of food photos. This research uses the composition feature extraction method using the SAMP-Net module, and the texture feature extraction method using Gray Level Co-occurrence Matrix (GLCM) and wavelet transform. Experiments were conducted on the public Gourmet Photography Dataset consisting of 24,000 food photos. The classification models used were Support Vector Machine (SVM), Random Forest, Multilayer Perceptron (MLP), and VGG16+MLP. The results were evaluated using metrics such as accuracy, precision, recall, and f1-score for each classification model used. From the feature ablation results, it was found that the GLCM texture feature had the most positive influence on the accuracy of the classification model. Among the four classification models used, the VGG16+MLP model has the best performance with an accuracy of 0.9108, precision of 0.9072, recall of 0.9321, and f1-score of 0.9195.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	GLCM, klasifikasi, komposisi, tekstur, wavelet, classification, composition, GLCM, texture, wavelet
Subjects:	Q Science > QA Mathematics > QA336 Artificial Intelligence
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Haniif Ahmad Jauhari
Date Deposited:	01 Aug 2024 13:49
Last Modified:	01 Aug 2024 13:49
URI:	http://repository.its.ac.id/id/eprint/110126

Actions (login required)

View Item