Abstract / Description of output
This chapter discusses the nature of evaluation in the context of automated solutions to map generalization and highlight the challenges of defining evaluation criteria. Considerable progress has been made in defining and incorporating evaluation criteria in systems that generalize maps over moderately small changes in scale. It gives an idealized view of how evaluation methodologies might be incorporated into generalization systems. For larger changes in scale, it will be necessary to evolve these techniques, and incorporate evaluation techniques linked to pattern analysis and a more synoptic evaluation of the success of the design solution overall. The chapter highlights the challenges of developing and prioritizing evaluation criteria, and where in the process they can be applied. Given the complexity of the task it is important to consider the role of the human in the evaluation process. Generalization techniques make changes to the map within a competing set of design goals. Consequently, a very broad set of techniques are required to measure and evaluate each of these changes to ensure optimal compromise within those design constraints. Evaluation is critical to the development of autonomous systems. The absence of more complete evaluation frameworks in which visual characterization and evaluation can take place remains as a major stumbling block to more fully automated map generalization solutions.