🤖 Internet Of Things (CCF 351)

The motion classifier classifies objects into static and dynamic, threby suppressing falsely predicted trajectories of static objects.

Arduino based table tennis shot predictor. Prediction is made in terms of probability, where the 4 type of gestures are assigned a probability. It can also be viewed as an accuracy score on your intended gesture if your training data is accurate enough.

It's classify the motion and vibration data from a stand fan. This is to mimic using embedded machine learning in an industrial environment. We want to be able to determine if a machine is off, on, low, high, anomaly, etc.

Jupyter notebook for classifying motion capture activities (Boning vs. Slicing) using feature engineering and machine learning. Includes data preprocessing, composite feature creation, feature selection, SVM and Random Forest model training, hyperparameter tuning, PCA, and model comparison.

2024 캡스턴 프로젝트 개발 저장소 with Flutter and Time Series Deeplearning