I'll develop a mobile app for recognizing dish calories from photos

I work on developing mobile apps that turn a plate photo into a ready nutrition report: the user points the camera, the app itself recognizes the dish, assesses portion size, and calculates calories and macros — without manually searching for a product in a database and entering grams by hand. This is a different approach than typical nutrition trackers where all data entry work falls on the user: here the camera makes the first draft, and the user only fixes inaccuracies.

I start with a combination of a computer vision model that segments the plate and identifies likely ingredients, and a calorie calculator tuned to real product composition tables. I pay special attention to the manual correction screen: you can change portion size, replace an ingredient, or add a missed one. A confirmed meal entry goes into the diary, and I make the monthly history with a calendar view and dynamic charts.

I take the project turnkey — an app, a server for photo processing and history storage, an admin panel for updating the recognition database. The work typically includes:
— recognition with model confidence indicator
— manual correction and manual ingredient search
— calorie and macro breakdown by meal
— monthly history with calendar and charts
— push reminders and onboarding for new users