image_processing.py

# -*- coding: utf-8 -*-

import numpy as np 

# functions to process captcha images

def bfs(visited, queue, array, node):
    # I make BFS itterative instead of recursive
    def getNeighboor(array, node):
        neighboors = []
        if node[0]+1<array.shape[0]:
            if array[node[0]+1,node[1]] == 0:
                neighboors.append((node[0]+1,node[1]))
        if node[0]-1>0:
            if array[node[0]-1,node[1]] == 0:
                neighboors.append((node[0]-1,node[1]))
        if node[1]+1<array.shape[1]:
            if array[node[0],node[1]+1] == 0:
                neighboors.append((node[0],node[1]+1))
        if node[1]-1>0:
            if array[node[0],node[1]-1] == 0:
                neighboors.append((node[0],node[1]-1))
        return neighboors

    queue.append(node)
    visited.add(node)

    while queue:
        current_node = queue.pop(0)
        for neighboor in getNeighboor(array, current_node):
            if neighboor not in visited:
                visited.add(neighboor)
                queue.append(neighboor)

def removeIsland(img_arr, threshold):
    # !important: the black pixel is 0 and white pixel is 1
    while 0 in img_arr:
        x,y = np.where(img_arr == 0)
        point = (x[0],y[0])
        visited = set()
        queue = []
        bfs(visited, queue, img_arr, point)
        
        if len(visited) <= threshold:
            for i in visited:
                img_arr[i[0],i[1]] = 1
        else:
            # if the cluster is larger than threshold (i.e is the text), 
            # we convert it to a temporary value of 2 to mark that we 
            # have visited it. 
            for i in visited:
                img_arr[i[0],i[1]] = 2
                
    img_arr = np.where(img_arr==2, 0, img_arr)
    return img_arr