feat: generating kernel

deblur/symetric_kernel.py

@@ -0,0 +1,35 @@
+import numpy as np
+
+
+def generate_kernel(radius, sigma=None):
+    """
+    Generate a 2D Gaussian kernel with a given radius.
+
+    Parameters:
+    - radius: int, the radius of the kernel (size will be 2*radius + 1)
+    - sigma: float (optional), standard deviation of the Gaussian. If None, sigma = radius / 3
+
+    Returns:
+    - kernel: 2D numpy array of shape (2*radius+1, 2*radius+1)
+    """
+    size = 2 * radius + 1
+    if sigma is None:
+        sigma = radius / 3.0  # Common default choice
+
+    print(f"radius: {radius}, sigma: {sigma}")
+
+    # Create a grid of (x,y) coordinates
+    ax = np.arange(-radius, radius + 1)
+    xx, yy = np.meshgrid(ax, ax)
+
+    # Apply the 2D Gaussian formula
+    kernel = np.exp(-(xx**2 + yy**2) / (2 * sigma**2))
+    kernel /= 2 * np.pi * sigma**2  # Normalize based on Gaussian PDF
+    kernel /= kernel.sum()          # Normalize to sum to 1
+
+    return kernel
+
+
+if __name__ == "__main__":
+    kernel = generate_kernel(radius=10, sigma=1)
+    print(kernel)