-
-
Application in Action
Inspiration :
The motivation behind creating this solution stems from a personal experience with my mother, who has diabetes. We often found ourselves perplexed about what foods would be suitable for her condition. Witnessing her struggle inspired me to explore ways to simplify the process of determining glycemic indexes, hence the creation of the project.
What it does :
The application is designed to process images of various food items, including both prepared dishes like noodles and pasta, as well as fruits like oranges and mangoes. Initially, it identifies the type of food presented. Following this, the app calculates the food's glycemic index. Based on the glycemic index result, the application then advises whether the food is suitable for consumption by individuals with diabetes and, if so, suggests an appropriate serving size.
How I built it :
The complete system was developed with Python and leverages the Gemini Vision Pro model API provided by Google. For the application's user interface, the open-source UI library Gradio was utilized.
Challenges I ran into :
The most significant challenge I faced was not in the accuracy of the Gemini Vision Pro model, which performed exceptionally well in my tests, but rather in conceptualizing and integrating the entire solution into a seamless user experience. The process of designing an intuitive interface and ensuring the app could handle a variety of food items with different textures, shapes, and sizes required a considerable amount of effort and iteration. Ensuring that the app could provide accurate glycemic index values in real-world scenarios was also a complex task that involved extensive testing and refinement.
Accomplishments that I am proud of :
One of my proudest accomplishments is the fact that the app has become a part of my mother's daily routine, helping her make informed dietary choices with ease. It's incredibly rewarding to see a personal project have such a positive impact on a loved one's life. Additionally, the app has gained significant traction within the community, with numerous users relying on it for their dietary management. The positive feedback and the growing user base are testaments to the app's value and effectiveness. Furthermore, the explanatory video I created to showcase the solution has been viewed over 15,000 times, indicating a strong interest and need for such a tool in the market.
What I learned :
Throughout the development of this app, I learned a great deal about the intricacies of dietary management for diabetes patients. I also gained insights into the capabilities and limitations of machine learning models when applied to real-world problems. The project honed my skills in API integration, user interface design, and the importance of user feedback in the iterative design process. I learned that creating a health-related app comes with a unique set of responsibilities, including ensuring accuracy, reliability, and user trust.
What's next for Diabetes Management App With Google's Gemini Vision Pro :
Moving forward, I plan to expand the app's reach by developing dedicated Android and iOS versions, making it accessible to a wider audience. I aim to incorporate user feedback to refine the app's features and improve its usability. Additionally, I will explore partnerships with health professionals to validate the app's recommendations and potentially integrate personalized dietary advice. My goal is to establish the app as a trusted companion for individuals managing diabetes and to continue enhancing its capabilities to support healthy eating habits.
Log in or sign up for Devpost to join the conversation.