Abstract
Human observers are able to group elements and fragments into whole shapes. This is the underlying concept of emergence. Yet, it is still unclear how information is sampled when emergence occurs. Usually, the process of object recognition is too fast to trace, but by using images with emerging properties we were able to study eye movements before, during and after the moment of recognition. Stimuli consisted of computer-generated emerging images (Mitra, Chu, Lee, and Wolf (2009)) resembling the famous Dalmatian against a dappled background. Forty observers each viewed fifteen emerging images while their eye movements were tracked. They indicated the moment of recognition by pressing a key. Results show that three phases could be distinguished in the eye movement patterns prior to and after recognition. In the first phase, initial fixation durations were relatively short and eye movements were distributed over the entire stimulus. In the second phase, which started approximately one second before recognition, there was a transition towards longer fixations targeted at the edges of the object. In the third phase, which started following recognition, fixations again became shorter and somewhat more distributed. Within the second phase just prior to recognition, additional viewing patterns could be identified: fixations were first made just outside the edges of the object, then there were more fixations inside the object and finally, fixations were made right on the edges of the emerging object. Our results show that the characteristics of eye movements are closely associated with the conscious experience of emergence. The distinct systematic patterns in the eye movements suggest that the recognition process can be divided into perceptually relevant temporal phases.
Meeting abstract presented at VSS 2014