弯曲图像的算法

Question

这是我试图实现的结果：

我的算法目前是：

1. 此画布照片在设计的位置是透明的。
2. 我调整了设计的大小，使其比帆布照片的大小略大一些
3. 我将画布照片粘贴到设计
上

它工作正常，但我需要图像在角落弯曲。有没有办法实现这个目标？

这是我使用PIL在python中编写的当前算法代码：

def merge_case_image(case_path, design_path, save_path):
    case = Image.open(case_path)  # cas template

    design = Image.open(design_path)  # design
    design.thumbnail(case.size, Image.ANTIALIAS)  # resize design
    design_w, design_h = design.size

    bg_white = Image.new('RGBA', case.size, (255, 255, 255, 255))  # white background
    bg_w, bg_h = bg_white.size

    offset = ((bg_w - design_w) / 2, (bg_h - design_h) / 2)

    if design.mode == "RGBA":
        bg_white.paste(design, offset, design)
    elif design.mode == "RGB":
        resizeImageMask = Image.new("L", design.size)
        resizedImagePixels = design.load()
        width, height = design.size
        for x in range(width):
            for y in range(height):
                cpixel = resizedImagePixels[x, y]
                if ((cpixel[0] == 255) and (cpixel[1] == 255) and (cpixel[2] == 255)):
                    resizeImageMask.putpixel((x, y), 0)
                else:
                    resizeImageMask.putpixel((x, y), 255)

        bg_white.paste(design, offset, resizeImageMask)

    bg_white.paste(case, (0, 0, bg_w, bg_h), case)

    bg_white.save(save_path, "PNG")