Neraca Dapur Cerdas dengan Image Classification dan Fitur Calorie Tracking

Rahadi, Gde Rio Aryaputra (2025) Neraca Dapur Cerdas dengan Image Classification dan Fitur Calorie Tracking. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5027201063-Undergraduate_Thesis.pdf Restricted to Repository staff only until 1 April 2027. Download (4MB) | Request a copy

Abstract

Pola hidup sehat menjadi salah satu gaya hidup yang banyak diminati oleh masyarakat modern, khususnya generasi muda yang terpapar oleh informasi dari media sosial. Salah satu aspek penting dalam pola hidup sehat adalah memantau dan mengelola asupan kalori dan nutrisi harian. Namun, metode konvensional yang digunakan untuk melakukan hal tersebut dinilai kurang praktis dan efisien, sehingga dapat mengurangi motivasi dan konsistensi dalam mencapai target kebugaran. Neraca dapur cerdas yang dikembangkan pada penelitian ini dapat mendeteksi jenis makanan dengan bantuan model Machine Learning yang melakukan klasifikasi citra, mengukur berat makanan dengan bantuan sensor load cell dan modul HX711, serta mengolah data jenis dan berat makanan untuk mendapatkan data kalori yang kemudian ditampilkan di layar LCD dan disimpan di aplikasi pengguna. Sistem berhasil dikembangkan dengan mengintegrasikan komponen Raspberry Pi 3, kamera OV5647 5MP, dan model custom MobileNet yang dilatih menggunakan platform Teachable Machine. Pengujian menunjukkan model dapat mengklasifikasikan dengan akurat (100%) untuk kelas “Chicken Breast”, “Banana”, dan “White Rice” pada berbagai kondisi, serta kelas “Tempeh” pada pengujian dengan background non-reflektif. Namun, performa klasifikasi menurun untuk sub-kelas tempe potongan dadu pada background reflektif dan pencahayaan tinggi. Data kalori tersimpan dalam format CSV dan terintegrasi dengan Blynk cloud untuk pemantauan jarak jauh. Tugas akhir ini diharapkan dapat memberikan manfaat bagi pengguna dari berbagai kalangan dalam melakukan calorie tracking dengan lebih praktis dan efisien.
=================================================================================================================================
Healthy lifestyle is becoming one of the most popular lifestyles among modern society, especially the young generation who are exposed to information from social media. One of the important aspects of healthy lifestyle is to monitor and manage daily calorie and nutrient intake. However, the conventional method used to do so is considered less practical and efficient, potentially lowering motivation and consistency in achieving fitness goals. The smart kitchen scale developed in this final project is capable of detecting food types using a Machine Learning model for image classification, measuring food weight with the aid of a load cell sensor and an HX711 module, and processing food type and weight data to calculate calorie information. This data is then displayed on an LCD screen and stored in a user application. The system was successfully developed by integrating a Raspberry Pi 3, an OV5647 5MP camera, and a custom MobileNet model trained using the Teachable Machine platform. Testing demonstrated that the model accurately classified various food classes with 100% accuracy such as the "Chicken Breast," "Banana," and "White Rice" classes under various conditions, as well as "Tempeh" class when tested against a non-reflective background. However, classification performance declined for diced Tempeh on reflective backgrounds and under high lighting conditions. Calorie data is stored in CSV format and integrated with the Blynk cloud for remote monitoring. This final project is expected to benefit users from various backgrounds in performing calorie tracking more practically and efficiently.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Calorie tracking, Convolutional Neural Network, Internet of Things (IoT), Machine Learning, Raspberry Pi
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7882.P3 Pattern recognition systems
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis
Depositing User:	Gde Rio Aryaputra Rahadi
Date Deposited:	22 Jan 2025 07:42
Last Modified:	22 Jan 2025 07:42
URI:	http://repository.its.ac.id/id/eprint/116602

Actions (login required)

View Item