Online separation of handwriting from freehand drawing using extreme learning machines

摘要: Online separation between handwriting and freehand drawing is still an active research area in the field of sketch-based interfaces. In last years, most approaches this have been focused on use statistical methods, which achieved significant results terms performance. More recently, Machine Learning (ML) techniques proven to be even more effective by treating problem like a classification task. Despite this, also these several aspects can considered open problems, including: 1) trade-off performance training time; 2) from different types drawings. To address just reported drawbacks, paper novel algorithm based set original features Extreme (ELM) proposed. Extensive experiments wide range sketched schemes (i.e., text graphical symbols), numerous than those usually tested any key work current literature, highlighted effectiveness proposed approach. Finally, measurements accuracy speed computation, during both testing stages, shown that ELM considered, area, better choice if compared with other popular ML techniques.