Experimenting accuracy and effectiveness in photo-interpretation for rapid crisis response and damage assessment applications through gaze control

摘要

The main objective of Geographic Information and Visual Perception project (GI&VP), and of the present study, was to offer a device composed of a suite of tools and based on gaze integration to support the work of photo-interpreters involved in remotely sensed imagery exploration for rapid crisis response and damage assessment applications. The system based on eye tracking technology was conceived to improve the interpreter accuracy and effectiveness. Our research challenge was to test the accuracy and the effectiveness during a non-standard photo-interpretation procedure: in the case of accuracy, we were interested in the number of clicks and in their correct assignment; in the case of effectiveness, we evaluated the interpreter’s ocular behavior and the actions performed using the tools at his/her disposal, designed in the prototype. This experiment, and the theoretic perspective in which it was conceived, expressed the attempt to carry out a collaborative work scenario combining human and computer cooperation in order to minimize each others limitations. Therefore, an important research challenge was trying to assess how much the human relies on the computer automatic extraction, and, on the other hand, how much the computer can be a support to the human.