Close
Calori, our groundbreaking software application, utilizes artificial intelligence to effortlessly determine the calorie content of meals – a potent tool in the fight against obesity, all at the tap of a button.
Calori’s genesis lies in the global obesity epidemic, impacting around 700 million adults. Beyond this, we confront the broader issue of malnutrition, addressing stunting, wasting, and obesity. Shockingly, the South Asian region, including Sri Lanka, bears the brunt with a staggering 14.1% prevalence of stunting among children under 5.
Calori’s relevance transcends borders, contributing to the global mission of reducing malnutrition. By honing in on energy-dense foods, our application empowers users to take control of their diets, promoting healthier living.
Crafting Calori involved integrating cutting-edge technologies and employing meticulous methodology. The app’s foundation rests on a Mask Region-based Neural Network (M-RCNN), trained for 7 epochs with a non-repetitive test-train split of 30 to 80 images of Sri Lankan food. While the current accuracy stands at 63.6%, we acknowledge its potential enhancement through increased training data and epochs.
Our development process employed a diverse toolset, with the Godot Game Engine for UI and graphics, Python for ML scripts, and Google Collaboratory for neural network training. Beyond calorie detection, the app incorporates an interactive daily-calorie intake calculator and BMI measurement, providing users with comprehensive health insights.
