Increasing efficiency of road extraction by self-diagnosis

In general, automatic object extraction systems may not be expected to deliver absolutely perfect results, and thus, for meeting predefined application requirements, a human operator must inspect the automatically-obtained results. In order to speed up the time- and cost-intensive inspection, the system should provide the operator with confidence values characterizing its own performance. In practice, however, this is rarely the case. In this paper, we present general ideas on the methodology and representation of internal evaluation (self-diagnosis) within automatic object extraction systems. We illustrate their implementation into two different road extraction systems, and bused on a test series of aerial images, we exemplify how results attached with confidence values can increase system efficiency for practical applications. To analyze the reliability of self-diagnosis, we matched the internally-evaluated results to a manually-plotted reference. The comparison shows the benefits but also some remaining deficiencies of the self-diagnosis tool.