🔬 ResearchAdded June 2026
Real-Time Filipino Sign Language Recognition Using MediaPipe
Undergraduate research paper (2nd author). We trained a lightweight gesture-recognition model on 4,500 samples of 30 common FSL signs, achieving 91.2% accuracy on a laptop webcam with no GPU. Presented at the 2025 PUP College of Computing Research Colloquium.
💭 Reflection
My first taste of real research. I learned to read papers properly, to be honest about our limitations section, and that collecting good data is 80% of the work.
Skills demonstrated
- Python
- TensorFlow
- MediaPipe
- Data Collection
- Academic Writing