PharmaScan

Techstack

Video Demo

pharm.1.mp4

Inspiration

PharmaScan was inspired by the real-world challenges faced by one of our teammate's aunts, who works in a pharmacy. She relies solely on pill colors to differentiate medications when preparing prescriptions, which can lead to errors and inefficiencies. Additionally, the physical and mental fatigue from long work hours highlighted the need for a more efficient and reliable solution. PharmaScan aims to reduce human error, enhance accuracy, and support pharmacy professionals in their daily tasks.

What It Does

PharmaScan is a smart medication identification and management tool that helps pharmacists streamline their workflow by:

Identifying pills based on physical characteristics such as color, and shape via ChatGPT-4o and imprinting them using image recognition. This provides essential quality assurance that the correct medication is processed and that it is the right amount. Our app reduces the reliance of pharmacists on manual identification to minimize fatigue and errors.

How We Built It

Frontend: React + TailwindCSS for an intuitive user interface.
Backend: Flask to handle data processing and API calls.
Machine Learning: Yolov8 for Computer Vision and Chatgpt-4o for medication matching
Database: Firebase for a secure authentication and Supabase + AWS S3 for Image Upload

We followed Agile development practices, allowing us to iterate quickly and respond to feedback effectively.

Information Architecture

Challenges We Ran Into

During development, we encountered several challenges:

1. Logistics Issues: We faced issues in creating our submission which eventually led to us not being able to submit to Devpost for IrvineHacks.
2. Data Availability: Finding comprehensive datasets for pill identification required manual verification and supplementation.
3. Model Accuracy: Training the AI model for high accuracy requires extensive tuning and data augmentation.
4. Time Constraints: Balancing feature development, testing, and deployment within tight deadlines was challenging.

Accomplishments That We're Proud Of

We're proud of several key achievements, including:

Successfully building a functional prototype that accurately identifies medications. Overcoming technical challenges and expanding our skill set with new technologies. Working effectively as a team to push through obstacles and deliver a working product especially since we mostly worked on the project during Saturday. We also received positive feedback from other competitors from judges and other hackers on our Project

What We Learned

Throughout the development of PharmaScan, we learned:

• The importance of user-centered design in creating solutions that address real-world problems.
• How to integrate machine learning models effectively into web applications.
• The complexities of handling healthcare data and ensuring accuracy and compliance.
• SUBMITTT THE APP TO DEVPOST BEFOREHAND!!!!!

What's Next

Though we didn't submit anything into the competition, we are happy to say that we developed an amazing product in our next iteration we are planning to:

• Develop an AI model for pill identification instead of utilizing ChatGPT-4o
• Revamp with a fresh UI and integrate Framer Motion
• Add another feature that correlates to ensuring everything fits
• Submit to another Hackathon!