Protractor: a fast and accurate gesture recognizer

Yang Li. 2010. Protractor: a fast and accurate gesture recognizer. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI '10). ACM, New York, NY, USA, 2169-2172. DOI=10.1145/1753326.1753654 http://doi.acm.org/10.1145/1753326.1753654


This paper discusses Protractor, a template-based gesture recognizer focused on low memory requirements and fast classification. Protractor uses a nearest neighbor approach, learning from user input training data. The gesture can be specified to be sensitive to orientation, or invariant. The preprocessing of the gesture is similar to the $1 recognizer, with 16 points total used. Classification is accomplished using the optimal angular distances (inverse cosine distance between template and sample vector values). Protractor also uses a closed form solution to find a rotation of the gesture which leads to the least distance. Protractor was compared to the $1 recognizer and was found to perform similarly, but with slightly faster response times.

Protractor is a promising contribution for mobile devices constrained for processing power and memory. It offers a simple algorithm for user defined gestures.

"Those look similar!" issues in automating gesture design advice

A. Chris Long, James A. Landay, and Lawrence A. Rowe. 2001. "Those look similar!" issues in automating gesture design advice. In Proceedings of the 2001 workshop on Perceptive user interfaces (PUI '01). ACM, New York, NY, USA, 1-5. DOI=10.1145/971478.971510 http://doi.acm.org/10.1145/971478.971510

This paper discusses developing a system for application designers to use for judging the similarity of gestures. The paper claims that both human perceptual similarity and computer recognition similarity are considered in advising the designer. The human perceptual data is based on a very limited sampling of which gestures people believed to be similar or different. The machine recognition similarity is based on how well the gesture is distinguished using the Rubine algorithm.

Overall this paper's results are questionable at best. The paper makes several contradictory claims and uses a limited sample size as its basis. Furthermore, the authors follow their "intuition" rather than designing rigorous studies, and admit that their system is incorrect in several cases. Much of the paper is spent describing basic HCI which is not related to gestures. The paper offers very few technical descriptions. In summary, this paper is of limited use for modern gesture design.