How to overlay images with transparent background?-CodePudding

Let say that I have a 2 pictures that are transparent. How can I overlay picture 1 to picture 2 so to get a result picture. Picture 1 is also smaller that picture 2. I assume that it can be done with opencv or PIL (GIMP is not allowed)

Picture number 1 :

And picture number 2 :

[

That's the result that I want to get:

My approach : I thought that I can be done with opencv function cv.addWeighted(src1, alpha, src2, beta, gamma[, dst[, dtype]]) what I also tried to perform. But as the matter in fact major problems are next points: pictures are transparent and the task is to leave pictures with higher resolution. Rely on your answers.

CodePudding user response：

As a matter in fact, I find out the solution that suits for me. Perhaps It's not an optimal solution but at least It works.

The main idea is that I merge(paste) 2 pictures into 1 result using PIL library and the method paste, so that I have background and foreground.

In my case I have Picture 1 - foreground and Picture 2 - background.

My code :

# This function fit picture 2(background) for picture 1 . It's pare
def transform (template):
    
    img = Image.open(template)
    img_w, img_h = img.size
    image = img.resize((img_w coef_w,img_h coef_h),Image.ANTIALIAS)
    # change coef_w and coef_h to resize to fit template 
    image.save('updated_picture.png',"PNG")

# This function merge,compose, concat) pictures 
def compose (image, template):
    img = Image.open(image)
    img_w, img_h = img.size
    print(img.size)
    background = Image.open(template)
    print(background.size)
    # resize the image
    size = background.size
    background = background.resize(size, Image.ANTIALIAS)
    bg_w, bg_h = background.size
    offset = ((bg_w - img_w) // 2, (bg_h - img_h) // 2)
    background.paste(img, offset, mask=img)
    background.save(f'{image}_final.png', "PNG")

How it's work, I transform template(ears) so they would lay on the picture at center and the next one script offset foreground by center so I can get next result.

Transform is needed to make template smaller or bigger. The results would depend on of it's template. I mean the scale of ears in my case. At the next picture you can see what I mean