Performance evaluation of line simplification algorithms for vector generalization

Abstract

Many studies of line simplification methods have been developed; however, an evaluation of these methods is still an open issue. This paper aims to evaluate a diversity of automatic line simplification algorithms in terms of positional accuracy and processing time. Past research studies for the performance evaluation were centred on measuring the location difference between a line to be simplified and its simplified version. However, the original line contains positional uncertainty. This paper evaluates performance of the line simplification algorithms using two comprehensive measures of positional accuracy of the simplified line. These two measures include one displacement measure and one shape distortion measure, both of which are able to consider (a) the displacement between the original line and its simplified version, and (b) positional uncertainty of the original line.