What!?! No Rubine Features?: Using Geometric-based Features to Produce Normalized Confidence Values for Sketch Recognition

As pen-based interfaces become more popular in to- day's applications, the need for algorithms to accu- rately recognize hand-drawn sketches and shapes has increased. In many cases, complex shapes can be constructed hierarchically as a combination of smaller primitive shapes meeting certain geometric constraints. However, in order to construct higher level shapes, it is imperative to accurately recognize the lower-level prim- itives. Two approaches have become widespread in the sketch recognition field for recognizing lower-level prim- itives: gesture-based recognition and geometric-based recognition. Our goal is to use a hybrid approach that combines features from both traditional gesture- based recognition systems and geometric-based recogni- tion systems. In this paper, we show that we can pro- duce a system with high recognition rates while provid- ing the added benefit of being able to produce normal- ized confidence values for alternative interpretations; something most geometric-based recognizers lack. More significantly, results from feature subset selection indi- cate that geometric features aid the recognition process more than gesture-based features when given naturally sketched data.