70 likes | 176 Vues
Explore the accuracy of the multilevel perceptron algorithm in Weka for classifying side kicks to develop a beta training application for beginners. Capture data on side kicks of various types using skeleton joint angles and experiment with different algorithms in Weka. Overcome challenges related to skeleton recognition and class complexity. Share your tips, thoughts, and questions on this machine learning project.
E N D
Project Proposal Machine learning with Weka for Kung fu-training purposes. Victoria Værnø
With what accuracy can the multilevel perceptron algorithm in Wekaclassify side-kicks? Is it good enough for a beta training application for complete beginners?
Data Capture 500 side kicks of types: ”Bad Kick” ”Good Kick Guard Down” ”Good Kick Guard Up” Attributes: Skeleton joint angles. Data Processing • Reducemoviesizesto 6 frames. • Make ARFF file with all the kicks. • Feedthe ARFF file intoWeka and experimentwiththealgorithms.
Challenges • Skeletonrecognition • Complexclasses • Too muchsimilaritywhich is irrelevant to theclassification
Questions? Tips? Thoughts?