hukuto7750
hukuto7750

Reputation: 149

How To Random Crop (Specific Area And Specific Probability) Rectangle From Image

This is my now code (it could crop image to 25 piece jpg from one image)

# -*- coding:utf-8 -*-
from PIL import Image

def cut(id,vx,vy):
    # 打開圖片(open image)
    name1 = "C:\\Users\\admin\\Desktop\\normal_random_crop\\test.png"
    name2 = "C:\\Users\\admin\\Desktop\\normal_random_crop\\test_" + id + "_crop.jpg"
    im =Image.open(name1)
    #偏移量(offset)
    dx = 100
    dy = 100
    n = 1
    #左上角切割(Left Top Poit)
    x1 = 0
    y1 = 0
    x2 = vx
    y2 = vy
    #縱向(Vertical)
    while x2 <= 512:
        #橫向切(Horizontal)
        while y2 <= 512:
            name3 = name2 + str(n) + ".jpg"
            im2 = im.crop((y1, x1, y2, x2))
            im2.save(name3)
            y1 = y1 + dy
            y2 = y1 + vy
            n = n + 1
        x1 = x1 + dx
        x2 = x1 + vx
        y1 = 0
        y2 = vy
    return n-1

if __name__=="__main__":
    id = "1"
    #切割圖片的面積 vx,vy (Crop Area)
    res = cut(id,100,100)
    print(res)

i hope to make generate amount random crop and specific every area specific probability for example : random crop 100 piece (from 512x512 image) total 104%

(

=====1%+2%+1%=====

1%+10%+10%+10%+1%+

1%+10%+10%+10%+1%+

1%+10%+10%+10%+1%+

=====1%+2%+1%=====

)

yellow area delete(don't need)

enter image description here

Upvotes: 2

Views: 461

Answers (1)

khachik
khachik

Reputation: 28693

First of all, the total probability can't be more than 100% (in least in this universe). Assuming it is 100%, you can represent the image as 1-dimensional array and then do a weighted random choice.

So if you target 25 pieces, 5x5, with probabilities of

0 1  2  1  0
1 9  10 9  1
1 10 10 10 1
1 9  10 9  1
0 1  2  1  0

then it becomes a simple list of probabilities:

[0, 1,  2,  1,  0, 1, 9,  10, 9, ... # 25 elements total]

Then you can make a weighted random choice, choose you way, for example, from here: A weighted version of random.choice

Upvotes: 1

Related Questions