[FIXED] Python - How to Put Image as Picture Frame for Other Image

Issue

I have a list of URLs that contain nature images with different resolutions (JPEG/JPG).

I need to make a PNG image with an empty background as a responsive picture frame (full frame) for those images in Python.

I have tried to use Pillow, but it didn't work as expected, and it seems I also need to alter the Picture Frame resolution for each of those images for the image to fit-in.

import urllib.request
from PIL import Image

urllib.request.urlretrieve('https://images.unsplash.com/photo-1666030910636-6291b581962e?ixlib=rb-1.2.1&ixid=MnwxMjA3fDB8MHxwaG90by1wYWdlfHx8fGVufDB8fHx8&auto=format&fit=crop&w=1000&q=80', "nature1.jpg")
nature_image = Image.open("nature1.jpg")
picture_frame = Image.open("pict_frame.png")
picture_frame.paste(nature_image, (50,50))
picture_frame.show()

Is there a way to achieve the result I expected? Should I use OpenCV/other libraries?

Picture Frame

Image

Result

Expected Result

Solution

Assume you want to enlarge the frame to fit the picture. Then here is one way to do that in Python/OpenCV.

• Read the picture
• Read the frame with alpha
• Separate the base (bgr) image from the alpha
• Threshold the alpha channel to binary
• Apply morphology to close up any small gaps
• Get contours
• Find the smallest (interior picture area) contour
• Get its bounding box values
• Draw the contour on a copy of the frame to be sure it is correct
• Compute the scaling between the frame and picture
• Resize the frame to fit the picture in the contour region
• Add an opaque alpha channel to the picture
• Put the picture into the resized frame
• Save results

Image:

Frame:

import cv2
import numpy as np

# load image
img = cv2.imread('mountain_scene.png')
hh, ww = img.shape[:2]

# load frame with alpha channel
frame = cv2.imread('frame.png', cv2.IMREAD_UNCHANGED)

# separate BGR and A channels
bgr = frame[:,:,0:3]
alpha = frame[:,:,3]

# threshold alpha to binary
thresh = cv2.threshold(alpha, 0, 255, cv2.THRESH_BINARY+cv2.THRESH_OTSU)[1]

# apply morphology to clean up any small gaps
kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (3, 3))
morph = cv2.morphologyEx(thresh, cv2.MORPH_CLOSE, kernel)

# get contours and find smallest (interior picture area)
contours = cv2.findContours(morph, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)
contours = contours[0] if len(contours) == 2 else contours[1]
small_contour = min(contours, key=cv2.contourArea)
x,y,w,h = cv2.boundingRect(small_contour)

# draw contour on copy of frame in opaque red
contour = frame.copy()
cv2.drawContours(contour, [small_contour], 0, (0,0,255,255), 1)

# scale the frame to fit the picture
scale_x = ww/w
scale_y = hh/h
xoff = int(x*scale_x)
yoff = int(y*scale_y)
new_frame = cv2.resize(frame, (0,0), fx=scale_x, fy=scale_y, interpolation=cv2.INTER_LINEAR)

# add opaque alpha to picture
img2 = cv2.cvtColor(img, cv2.COLOR_BGR2BGRA)

# put picture in new_frame
result = new_frame.copy()
result[yoff:yoff+hh, xoff:xoff+ww] = img2

# save results
cv2.imwrite("frame_contour.png", contour)
cv2.imwrite("frame_with_picture.png", result)

# show result
cv2.imshow("contour", contour)
cv2.imshow("frame_with_picture", result)
cv2.waitKey(0)
cv2.destroyAllWindows()

Contour on frame:

Picure in frame:

Answered By - fmw42