Purchase this article with an account.
Isamu Motoyoshi, Frederick A A Kingdom; Energy-frequency analysis reveals orientation-opponent channels in human vision. Journal of Vision 2003;3(9):614. doi: 10.1167/3.9.614.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
It has been suggested that the representation of global form and texture is subserved by visual mechanisms that integrate the outputs of 1st-order orientation-selective filters. We here directly characterized the input-orientation characteristics of such mechanisms using linear-system analysis. The stimulus was band-pass noise whose Fourier-energy was sinusoidally modulated across orientation (i.e., not across space). Sensitivity for detecting Fourier-energy modulation was measured for various orientation frequencies (1–20 cycle/pi), stimulus sizes (2.3–18.6 deg), and carrier spatial frequencies (0.9–6.9 cpd). An inverse Fourier transform of the sensitivity data revealed a strong centre-surround antagonism across orientation (excitatory centre within 6–9 deg and inhibitory lobes at 15–20 deg), that increased with stimulus size far beyond the size of orientation-selective filters (more than 64 times the spatial wavelength of the noise carrier). These results demonstrate the existence of 2nd-order mechanisms that integrate and differentiate 1st-order orientation signals over a wide area of the visual field when processing texture. We term these mechanisms ‘orientation-opponent’ channels, analogous to the colour-opponent channels in colour vision. We applied a similar analysis to the spatial-frequency domain, using fractal (1/f) noise whose Fourier energy was modulated across spatial frequency (1–5 cycle/octave). However, we failed to obtain clear evidence for ‘frequency-opponent’ channels.
This PDF is available to Subscribers Only