Intrinsic 2D features as textons

Erhardt Barth, Christoph Zetzsche, and Ingo Rentschler

Abstract: We suggest that intrinsic-2D (i2D) features, computationally defined as the outputs of nonlinear operators which model the activity of end-stopped neurons, play a role in preattentive texture discrimination. We first show that for discriminable textures with identical power spectra the predictions of traditional models depend on the type of nonlinearity and fail for energy measures. We then argue that the concept of intrinsic dimensionality, and the existence of end-stopped neurons, can help to understand the role of the nonlinearities. Furthermore, we show examples where models without strong i2D selectivity fail to predict the correct ranking order of perceptual segregation. Our arguments regarding the importance of i2D-features resemble the arguments of Julesz and colleagues on behalf of textons like terminators and crossings. However, we provide a computational framework which identifies textons with the outputs of nonlinear operators which are selective to i2D features.

(pdf  0.23 MB) (PostScript 15.6 MB)