MealToGo

Training process/Data
Detection
Precision matrix

Inspiration

We live in a fast-paced world where people often compromise on their health due to time constraints. Cooking healthy meals can be time-consuming and daunting, leading many to opt for quick, unhealthy alternatives. Additionally, there's a lack of personalization in meal planning, making it challenging for individuals with specific dietary needs to find suitable recipes.

What it does

MealToGo addresses these challenges by providing a platform that offers quick and nutritious meal options tailored to the user's health profile. Our solution utilizes cutting-edge technology, including machine learning algorithms and nutritional expertise, to curate personalized recipes in minutes.

How we built it

To build it we first took about 100 pictures for each fruit that we included in our demo. We then used Roboflow to create bounding boxes around them and separate them into training, testing, and validating images. Then we used YOLOv8 and Roboflow to train our model and running it for 50 epochs seemed to be the sweet spot. Finally, we downloaded the model and executed it in PyCharm.

Challenges we ran into

We ran into many challenges, the first one being that we started out by using a pre-trained model which was very inaccurate. Another big problem we had was the time crunch because time was going by very fast.

Accomplishments that we're proud of

When we started developing our app, it wasn’t just about creating something useful—it was about building a strong team. As we welcomed new members with unique skills, we faced and overcame many challenges together. From solving technical issues to meeting tight deadlines, every problem strengthened our teamwork and resilience. This process taught us the value of collaboration and perseverance, leading us to a successful framework of a potential business and forging lasting relationships among team members.

What we learned

There are many things that we learned. One thing we learned was communication and how to effectively communicate our ideas and thoughts about the project. Another thing that we learned, specifically Elil, was Flet and how to use it. We also learned how to use Roboflow to create bounding boxes around images so that AI can do object detection in them.

What's next for MealToGo

For the future, MealToGo plans on adding even more cooler features into our app. With the advancement of technology, we plan on implementing this into wearable electronics. This would help gather real-time health data, such as activity levels and vital signs, for even more accurate health profiles and recommendations. We also plan for inclusivity in our app as well by expanding our recipe database to cater to a diverse range of dietary preferences, including vegan, vegetarian, gluten-free, and more. We also planning on making this go global and creating dishes tailored to people of different cultures