shakey |||

Prototype counting pills

I worked on a deep learning pill counting and identification system. The model is a quantized MaskRCNN model running on a raspberry pi in real time. I was able to minimize reflections, reduce glare and other unwanted imagery by using a black background during training data collection. I collected and annotated a dataset of about 60 rows using Labelbox.

detection pill

Up next CLI Tool for bldc motor Beetleapp
Latest posts QR Gate opener Beetleapp Prototype counting pills CLI Tool for bldc motor Tonal Unity Sim Motor controller design BLDC Notes/Cheat Sheet Designing custom motor controllers Autonomous robots LA Hackathon