Seeing structure through chaos

Research on motion perception has concentrated on recognizing regularities in images due to rigid motion. Rigid motion, however, is only one type of motion that occurs. There are also elastic motion, fluid motion, and turbulent flow. The authors use a Brownian fractal model of turbulent flow, demonstrate that turbulence can be recognized visually, and suggest how the model can be used to make inferences about the flowing fluid. The recognition procedure is implemented by testing for a fractal scale-invariant regularity across space and time using linear spatiotemporal bandpass filters. These filters are also used to extract optical flow.