Biosensor Machine Learning with Julia
summary
Describes extracting EMG from consumer biosensors via the open source BrainFlow, processing and training in Julia, and deploying a saved model with JLD2 to do live gesture predictions for myo‑style muscle sensors. Data collection uses two‑minute labeled recordings per gesture, sliced into windows; a random forest prototype reaches about 80–90% accuracy in prototypes, streams predictions live with occasional flicker, and is demonstrated controlling games and a bionic‑arm control pipeline.
AI-written summary · about
chapters (12)
- 0:00Biosensors for data science
- 1:55Consumer biosensors and apps
- 4:23Julia for biosensor ML
- 5:27Getting biosensor data
- 6:30Plotting biosensor signals
- 8:18Gesture classification from muscle data
- 11:05Julia DataFrames vs pandas
- 12:51Random forest in Julia
- 15:20Live gesture prediction demo
- 16:24Controlling games with muscle signals
- 17:27Starting with biosensors today
- 18:11Q&A
related talks
22mSeeing, Sensing and Recognizing Laban Movement Qualities
28mStrike a Pose - Gesture recognition in JavaScript w/ ML & Arduino
34mThe hand prosthesis that learns from the user
08mDesigning a Personalised Sensor Glove Using Deep-Learning
35mTranslating Brain Electrical Signals into Actions through Models...
03mPowered by TensorFlow: Creating a custom, machine learning-powered drumming arm
22mStrike a Pose: Gesture Recognition in JS with Machine Learning
26mBrainDrain Using Machine Learning and Brain Waves to Detect Errors in Human