Inspiration
we built LabLens for the MRG Laboratories challenge at YCP Hacks. we wanted to make it easier for labs to turn handwritten sample labels into digital records.
What it does
LabLens lets users upload pictures of lab labels, extracts information using OCR, and organizes it into a CSV file. Users can review and correct mistakes before downloading.
How we built it
We used Python, Streamlit, EasyOCR, OpenCV, Pandas, and SQLite. I relied heavily on ChatGPT and Claude to help write and debug the code.
Challenges we ran into
The biggest challenges were recognizing handwriting, fixing bugs with saving and deleting samples, and making sure corrected information appeared in the CSV.
Accomplishments that we're proud of
Learning how models are trained on datasets.
What we learned
I learned a lot about Python, OCR, databases, and debugging. This was my first time building something like this.
90% vibe-coded, 100% learning.
What's next for LabLens
I want to improve handwriting recognition, fix remaining bugs, and eventually automate the scanning process.