psychos.visual.synthetic.gabor_2d#

psychos.visual.synthetic.gabor_2d(size: Tuple[int, int] = (128, 128), spatial_frequency: float = 5.0, orientation: float = 45.0, phase: float = 0.0, sigma: float | None = 0.25, gamma: float = 1.0, contrast: float = 1.0, baseline: float = 0.0, center: Tuple[float, float] | None = None, rescale: Literal['none', 'normalize', 'clip'] | None = None, vmin: float = 0.0, vmax: float = 1.0, return_envelope: bool = False) → ndarray | Tuple[ndarray, ndarray][source]#

Generate a 2D Gabor patch as a numpy array (grayscale), with an optional Gaussian envelope.

Parameters:

  • size (Tuple[int, int]) – (height, width) of the output array in pixels.

  • spatial_frequency (float) – Cycles per entire image width. For orientation=0, this yields exactly spatial_frequency cycles from left to right.

  • orientation (float) – Gabor orientation in degrees, [0..360]. Note: The function rotates the carrier by (360 - orientation) degrees.

  • phase (float) – Fraction of 2*pi for the carrier phase, in [0..1].

  • sigma (Optional[float]) – Standard deviation of the Gaussian envelope, as a fraction of the minimum dimension. If None, no envelope is applied (i.e., the envelope is 1 everywhere).

  • gamma (float) – Aspect ratio of the Gaussian envelope (width vs. height). If 1.0, the envelope is circular; values < 1 or > 1 yield an elliptical envelope.

  • contrast (float) – Contrast (amplitude) of the sinusoidal carrier.

  • baseline (float) – Baseline offset intensity added to the Gabor.

  • center (Optional[Tuple[float, float]]) – (y_center, x_center) specifying the center in pixel coordinates. If None, defaults to the image center ((h/2), (w/2)).

  • rescale (Literal["none", "normalize", "clip"]) –

    Rescaling method:

    • ’none’: return the raw computed values.

    • ’normalize’: linearly scale the array so that its min maps to vmin and max to vmax.

    • ’clip’: clip values to the range [vmin, vmax].

  • vmin (float) – Minimum value for normalization or clipping. Ignored if rescale==’none’.

  • vmax (float) – Maximum value for normalization or clipping. Ignored if rescale==’none’.

  • return_envelope (bool) – If True, return a tuple (gabor, envelope), where envelope is the Gaussian envelope used (useful for an alpha channel). If sigma is None, envelope is an array of ones.

Returns:

If return_envelope is False, returns a 2D array containing the Gabor pattern. Otherwise, returns a tuple: (gabor, envelope).

Return type:

Union[np.ndarray, Tuple[np.ndarray, np.ndarray]]

Example

>>> import matplotlib.pyplot as plt
>>> gabor, env = gabor_2d(size=(256, 256), spatial_frequency=2.0, orientation=45.0,
...                       phase=0.25, sigma=0.25, gamma=1.0, contrast=1.0,
...                       baseline=0.0, rescale='normalize', vmin=-1.0, vmax=1.0,
...                       return_envelope=True)
>>> plt.figure(figsize=(10, 5))
>>> plt.subplot(1, 2, 1)
>>> plt.imshow(gabor, cmap='gray', origin='lower')
>>> plt.title("Gabor Patch")
>>> plt.subplot(1, 2, 2)
>>> plt.imshow(env, cmap='gray', origin='lower')
>>> plt.title("Envelope (Alpha Channel)")
>>> plt.show()
On this page

This Page