The K-Means algorithm has been extended on university with a function recognition algorithm compatible with LearningAPI. Additionally, to algorithm, also an application MouseGestureRecognition has be created. All required artefacts can be found here: https://github.com/UniversityOfAppliedSciencesFrankfurt/LearningApi/tree/KMeans-branch/LearningApi
My task is to :
- Analyze existing application, which basically capture mouse pointer movements and collects coordinates. Set of coordinates is learning set of mouse points as array of two-dimensional vectors.
- Analyze and document how feasible is recognition of signature (function) by using of two-dimensional mouse coordinates.
- Extend existing application by adding a time as a third component in the vector.
- Analyze and document algorithm with time approach and compare it with previous results.