Overview

KaloriKu is a computer vision-based mobile application aimed at estimating calorie content from images of Indonesian traditional foods. Utilizing deep learning models trained on a diverse dataset of local dishes, the system identifies food types and calculates approximate caloric values to support healthier dietary choices. This culturally tailored AI solution empowers users to maintain balanced nutrition while celebrating Indonesia’s rich culinary heritage, bridging technology and tradition for better health awareness.

Findings

Its featured features include:

🔍 Food detection using MobileNetV2 deep learning model
📷 Upload photos for automatic calorie estimation
🌐 Modern web with Laravel (RESTful API) as backend
🔐 Secure login/register system with Laravel authentication
📊 Model accuracy up to 98%

🔧 My Role:

• Leading the team project and arrange team's weekly meeting

• Collaborate in the AI ​​model training process (data augmentation, normalization, model optimization)

• Creating AI model until 92% accuracy using TensorFlow

• Deploy Website using Railway, an infrastructure platform

📈 Results and Achievements:

• The AI ​​model achieved 95.22% validation accuracy in classifying typical Indonesian food.

• The project was successfully presented well and received positive feedback from mentors and participants of Coding Camp 2025.