Meal Tracking:
Problem: Difficulty in monitoring and managing nutritional intake.
Solution: NutriCraft simplifies the process by allowing users to easily track and analyze the nutritional content of their meals through image recognition technology.
Recipe Generator:
Problem: Lack of inspiration or uncertainty about what to cook with available ingredients.
Solution: NutriCraft provides a solution by generating personalized recipes based on user preferences, time constraints, and expertise level, making meal planning more accessible and enjoyable.
Stats and Visualization:
Problem: Tracking and understanding daily nutritional data can be overwhelming.
Solution: NutriCraft offers visualizations like heatmaps, pie charts, and bar charts to help users easily interpret their data, fostering better awareness of their eating habits and progress toward health goals.
One of the challenges we ran into was reducing hallucinations. Often the LLM tended to identify ingredients that werent even there. This was minimised by adjusting LLM configuration with parameters like temperatures anbd Top-P Top-K. Also we ran into a bit of an issue while transferring images from frontend to backend even using base64 due to size limits which we overcame.
Tracks Applied (1)
Discussion