Combine blocks of different images and produce a new image

Question

I have six photographs. I changed them into block structure. Consider an image of size 200x200. 1. I converted into blocks of 10x10 so now I have 400 blocks in total each of siz

Answer 1

So, here's my approach to your problem. I rewrote parts of your code to get rid of all the lists, and solely working on NumPy arrays. Therefore, I store all images in a 4D array, and store all calculated "block means" in a 3D array. Finally, I use the found image_number array to generate some kind of "index array" by using OpenCV's resize method with INTER_AREA interpolation flag ("resampling using pixel area relation"). In doing so, the generation of your final image can be done very easily using NumPy's boolean array indexing.

Let's have a look at the following code:

import cv2
import numpy as np

# Read images in one single 4D array; resize to [200, 200]
nImages = 3
images = np.zeros((200, 200, 3, nImages), np.uint8)
images[:, :, :, 0] = cv2.resize(cv2.imread('U2Gmz.png', cv2.IMREAD_COLOR), (200, 200))
images[:, :, :, 1] = cv2.resize(cv2.imread('OZxf3.png', cv2.IMREAD_COLOR), (200, 200))
images[:, :, :, 2] = cv2.resize(cv2.imread('aISEB.png', cv2.IMREAD_COLOR), (200, 200))

# Calculate block means and store in one single 3D array
means = np.zeros((20, 20, nImages), np.uint8)
for im in range(nImages):
    arr = np.split(images[:, :, :, im], 20)
    arr = np.array([np.split(x, 20, 1) for x in arr])
    means[:, :, im] = np.reshape([arr[i][j].mean() for i in range(20) for j in range(20)], (20, 20))

# Determine block mean maximum over all images
result = np.max(means, axis=2)

# Determine index of block mean maximum over all images
image_number = np.argmax(means, axis=2)
print(image_number)

# Resize index array with "resampling using pixel area relation" to final image size
image_number_idx = cv2.resize(np.uint8(image_number), (200, 200), interpolation=cv2.INTER_AREA)

# Generate final image by boolean array indexing
final = np.zeros((200, 200, 3), np.uint8)
for im in range(nImages):
    idx = image_number_idx == im
    final[idx, :] = images[idx, :, im]

# Show images
cv2.imshow('image1', images[:, :, :, 0])
cv2.imshow('image2', images[:, :, :, 1])
cv2.imshow('image3', images[:, :, :, 2])
cv2.imshow('final', final)

cv2.waitKey(0)
cv2.destroyAllWindows()

These are the used images:

The image_number output gives this:

[[0 0 0 1 1 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0]
 [1 0 0 0 1 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0]
 [1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0]
 [0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1]
 [0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 1 1 1 0]
 [0 0 0 0 0 0 0 2 1 1 1 2 0 0 0 0 0 1 0 0]
 [0 0 0 0 0 0 0 2 1 0 0 2 2 2 0 0 0 1 1 0]
 [0 0 0 0 0 2 2 2 1 0 2 2 2 2 0 0 0 1 1 0]
 [0 0 0 0 0 2 2 2 0 0 0 0 2 2 2 0 0 0 0 0]
 [0 0 0 0 2 2 2 2 0 0 0 0 2 2 2 2 0 0 0 0]
 [0 0 0 0 2 0 2 2 0 0 0 0 2 0 0 0 0 0 0 0]
 [1 1 0 0 0 2 2 0 0 2 2 0 0 2 0 0 0 0 0 0]
 [1 1 0 0 2 2 2 0 2 2 2 2 1 2 2 2 2 0 2 1]
 [1 0 0 0 0 2 2 2 2 0 2 2 2 2 2 2 0 1 1 1]
 [1 1 1 0 0 2 2 2 1 1 1 2 2 2 2 0 0 1 1 0]
 [1 1 1 1 1 1 1 1 1 1 1 2 0 0 1 0 0 0 0 0]
 [1 1 1 1 1 1 0 1 1 1 1 1 0 2 0 0 0 0 0 0]
 [1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]
 [1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0]
 [0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1]]

And, the final image looks like this:

I hope, I understood your question correctly, and this is what you wanted to achieve.

My assumption is, that all input images have the same image dimensions, (200, 200) here. Otherwise, I couldn't think of a way to manage potentially varying block sizes, if just the "grid", (20, 20) here, is fixed.

Hope that helps!

EDIT: To read all jpg files from the given folder, you might use:

files = glob.glob('resized/*.jpg')

# Read images in one single 4D array; resize to [200, 200]
nImages = len(files)
images = np.zeros((200, 200, 3, nImages), np.uint8)
for im in range(nImages):
    images[:, :, :, im] = cv2.resize(cv2.imread(files[im], cv2.IMREAD_COLOR), (200, 200))