CALY turns a meal photo or a natural-language description into recognized foods, estimated portions, nutritional values, and personalized daily guidance — inside one connected mobile experience.
Traditional tracking asks people to separate every ingredient, weigh it, search a database, calculate nutrition, and repeat the entire process for every meal. That friction is exactly what CALY was built to remove.
Every meal can turn into a long sequence of manual decisions before it is ever recorded.
Separate every food item and ingredient.
Estimate or measure each portion manually.
Search nutritional databases item by item.
Calculate calories and macronutrients repeatedly.
CALY supports different situations without forcing users into one rigid input method.
CALY separates the foods on the plate, estimates each portion, calculates nutritional values, and records the meal.
Write naturally in Arabic, Egyptian dialect, Franco-Arabic, English, or mixed language. CALY structures the description into reliable food records.
Save meals you eat often and add them again in seconds — without repeating the full analysis workflow.
The interface is simple because the engineering behind it is not. Each meal moves through specialized AI, mathematical, and software layers.
Swipe to explore all systems
Three fine-tuned YOLO instance-segmentation models run together. A custom Mask-IoU ensemble removes duplicates, protects nested toppings, resolves conflicts, and fills missed regions.
CALY was evaluated across model performance, portion accuracy, language understanding, response time, and user experience.
Curated and optimized for segmentation.
International and Egyptian dishes.
YOLOv11 with smart augmentation.
Across 50 diverse test meals.
Natural food descriptions.
End-to-end structured output.
Measured over standard 4G/LTE.
Seven-day user acceptance test.
CALY's Flutter application brings onboarding, personalized plan creation, meal logging, daily tracking, meal management, and recommendations into one consistent interface.
Personalized onboardingAge, gender, height, weight, activity level, and goal.
Smart meal loggingCamera, natural-language description, and favorites.
Daily dashboardCalories, protein, carbohydrates, and fat progress.
Meal controlReview, edit, delete, and save meals as favorites.
Designed in FigmaBuilt with FlutterThe backend orchestrates every step — from a user's input to AI inference, nutrition calculations, secure storage, and the updated result shown inside the application.
Profile setup, meal capture, text logging, favorites, and the daily dashboard.
Validates requests and coordinates the complete nutrition workflow.
YOLO, Depth Anything V2, Qwen 2.5, and Gemini.
Identity, profiles, meals, nutrition records, and uploaded images.
Structured result returns to the app: detected foods, estimated grams, nutritional values, daily progress, and personalized guidance.
CALY is our graduation project for the 2025–2026 academic year, developed by six students from the Statistics & Computer Science Program, Department of Mathematics, Faculty of Science, Capital University.
Professor of Mathematical Statistics
Download the beta APK, create your personalized nutrition plan, and explore CALY's meal logging and daily tracking experience.
