John K. Tsotsos

The Different Kinds of Visual Recognition need Different Attentional Binding Strategies

John K. Tsotsos

Dept. of Computer Science & Engineering and

Centre for Vision Research

York University

Toronto, Ont. Canada

Abstract Many think visual attention needs an executive to allocate resources. Although the cortex exhibits substantial plasticity, dynamic allocation of neurons seems outside its capability. Suppose instead that the visual processing architecture is fixed, but can be 'tuned' to task needs.The only resource that can be allocated is time. How can this fixed structure of the visual cortex be used over periods of time longer than one feed-forward pass? Can the Selective Tuning model provide the answer? This presentation puts forward the proposal that by using multiple passes of the visual processing hierarchy, both bottom-up and top-down, and using task information to tune the processing prior to each pass, we can explain the different recognition behaviors that human vision exhibits. By examining in detail the basic computational infrastructure provided by the Selective Tuning model and using its functionality, four different kinds of binding processes are distinguished by tying them directly to specific recognition tasks and their time course. These are Convergence Binding for Detection and Categorization tasks, Partial Recurrence Binding for Identification tasks, Full Recurrence Binding for Localization Tasks and Iterative Recurrence Binding for more complex tasks. Each will be illustrated using examples from our model simulator. The framework presented offers the possibility of reconciling the myriad of apparently disparate results in the field by noting which task type and what time course each may involve. This also is a challenge for others to help refine this early and clearly incomplete decomposition.