Recent Studies on Mathematical Modeling of Visual Cortical Cells

Recent studies on mathematical modeling of visual cortical cells [Kulikowski/Marcelja/Bishop:1982] suggest a tuned band-pass filter bank structure. These filters exhibit Gaussian transfer functions in the frequency domain. Consequently, applying the inverse Fourier transform to these transfer functions yields filter characteristics closely resembling Gabor filters. Gabor filters essentially consist of a Gaussian function (with variances sx and sy along the x and y axes respectively) modulated by a complex sinusoidal wave (with center frequencies U and V along the x and y axes). In practical implementation, Gabor filters can be generated using mathematical operations where a 2D Gaussian envelope is multiplied by a complex exponential function representing orientation and frequency selectivity. Recent research demonstrates that mathematical models can effectively characterize visual cortical cell properties. These cells exhibit tuned band-pass filter bank structures with Gaussian transfer functions in the frequency domain. By applying inverse Fourier transforms to these transfer functions, we obtain filter characteristics analogous to Gabor filters. The Gabor filter fundamentally comprises a Gaussian function with different variances (sx and sy) along the x and y axes, multiplied by a complex sinusoidal wave with distinct center frequencies (U and V) along the same axes. Code implementation typically involves creating 2D Gaussian kernels and combining them with complex sinusoids through element-wise multiplication operations.