Human Perceptual Factors in Imaging: A Link Between Cognitive And Computer Sciences

For a long time computer science and psychology have been investigated as two completely different scientific fields. Nevertheless, the history of the two sciences has proved that only the interplay of the two made it possible to reach the majority of the most relevant findings in fields that are interdisciplinary by nature as human computer interaction, robotic haptics research, computer graphics, brain-computer interfaces, computer security and particularly computer vision where theories and observations of visual perception have been the main source of inspiration. Several examples, some of which are reported in this thesis, proved the mandatory demand of performing studies at the intersection between the computer science and the psychology. In this thesis we have the ambitious purpose to "build" a solid bridge between these two sciences by blending knowledge and methodology coming from both the fields. To prove the achievement of our purpose three specific case studies have been investigated: color naming, human brain mapping during simulated surgical tasks and social signal processing. Color vision and color naming are very important topics in cognitive psychology but it also have relevant importance to develop mathematical models for the Italian language and perception. In 1969 Berlin and Kay suggested the existence of a semantic universal- ism in color vocabular and one of the main problems that they identified in their research was how to define color categories. In this framework, three main questions were investigated: The color categorization, the existence of a twelfth basic color term, and the extraction of a metric for similarity between pairs of color were examined in the Italian language. In surgical robotics, extensive evidence has suggested the advantages of minimally invasive surgery over open surgery, but contradictory results concerning the benefits of 3D displays used during minimally invasive surgery have been reported. Moreover, the parietal cortex is known to be involved during the execution of tasks demanding spatial and attentional processes and, from the surgeons’s perspective, during surgical procedures both such processes are required. Nevertheless, few studies have investigated the role of the parietal cortex on the surgical performance. In order to elucidate this issue two main questions have been investigated: the involvement of the parietal cortex during visual- spatial and visuo-motor surgical tasks and the effect of monoscopic and stereoscopic vision in surgical robotics on the human performance. The automatic recognition of human behaviour for videosurveillance applications is, without doubt, one of the main challenges in the current multimedia framework. Despite that, the importance of the role of the human being for reaching an efficient intelligent monitoring of the human activity has frequently been underestimated. In this framework, the so-called Social Signal Processing aims at developing theories and algorithms codifying how human beings behave while involved in social interactions, putting together perspectives from sociology, psychology and computer science. Starting from this field three objectives were pursued in this thesis: the automatic detection of groups and the identification and inference of social relationships by the use of three different social cues namely gaze, F-formations and proxemics. Even if this dissertation mainly focuses on the investigation of color naming issues, the study of brain mapping in surgical robotics and the investigation of social signal processing issues provide clear evidence of the need for integrating such two disciplines for moving a step forward in modern computer applications.