I\'m complitely new to Flood Fill algorithm. I checked it out from Wikipedia (http://en.wikipedia.org/wiki/Flood_fill). But didn\'t become that much wiser. I\'m trying to us
There are several implementations of the flood fill algorithm in image processing libraries for Python. I'm aware of two: skimage.segmentation.flood and OpenCV's floodFill. The former is implemented in Python using an algorithm similar to the one in amit's answer above. The latter is implemented in C++ using a conceptually similar algorithm, but without recursion, making it much more efficient (about 25x for large images).
To use OpenCV's floodFill, you'd need to convert your matrix to an np.array of integers, which could be done as follows:
import numpy as np
import cv2
matrix_np = np.asarray(matrix)
numeric_matrix = np.where(matrix_np=="a", 255, 0).astype(np.uint8)
mask = np.zeros(np.asarray(numeric_matrix.shape)+2, dtype=np.uint8)
start_pt = (y,x)
if matrix_np[start_pt]:
cv2.floodFill(numeric_matrix, mask, start_pt, 255, flags=4)
mask = mask[1:-1, 1:-1]
matrix_np[mask==1] = "c"
matrix = matrix_np.tolist()
With the example matrix you gave above and x,y=(0,0), this will set matrix
to
[['c', 'c', 'b', 'a', 'a', 'b'],
['c', 'b', 'b', 'a', 'b', 'b'],
['b', 'a', 'b', 'a', 'a', 'b'],
['b', 'a', 'b', 'a', 'b', 'b'],
['a', 'a', 'b', 'a', 'a', 'a'],
['a', 'b', 'b', 'a', 'a', 'b']]
Well, the idea of flood fill is:
python-like pseudo code:
def floodfill(matrix, x, y):
#"hidden" stop clause - not reinvoking for "c" or "b", only for "a".
if matrix[x][y] == "a":
matrix[x][y] = "c"
#recursively invoke flood fill on all surrounding cells:
if x > 0:
floodfill(matrix,x-1,y)
if x < len(matrix[y]) - 1:
floodfill(matrix,x+1,y)
if y > 0:
floodfill(matrix,x,y-1)
if y < len(matrix) - 1:
floodfill(matrix,x,y+1)