Using U2 net deep network to realize -- certificate photo generation program

Results the preview ：http://map.gnnu.work/rm21/qy/profilepicture

Use to ：
1.pytorch Load split model
2.PIL More convenient to operate images
2.U2-net The network separates the contour of the portrait from the image

The specific implementation steps are as follows ：

1. The original picture

The original picture

2. Get from the original picture Alpha chart

#  Get from the original picture Alpha chart  org： Path of original graph   Return what you get Alpha chart 
def seg_trimap(org):
    image = Image.open(org)
    img = np.array(image)
    net = pre_net()
    inputs_test = pre_test_data(img)
    d1, d2, d3, d4, d5, d6, d7 = net(inputs_test)
    # normalization
    pred = d1[:, 0, :, :]
    pred = normPRED(pred)
    #  Convert data into pictures 
    im = get_im(pred)
    # im.save(alpha)
    sp = image.size
    #  Adjust the size according to the original picture 
    imo = im.resize((sp[0], sp[1]), resample=Image.BILINEAR)
    imo.save("alpha_resize111.jpg")
    return imo
    return imo

Alpha chart

3. take Alpha Turn the picture into trimap chart

def to_standard_trimap(alpha):
    # Alpha Graph generation  trimap
    # image = Image.open(alpha)
    image = alpha;
    # image = image.convert("P")
    # image_file.save('resize_trimap.png')
    sp = image.size
    width = sp[0]
    height = sp[1]
    for yh in range(height):
        for xw in range(width):
            dot = (xw, yh)
            color_d_arr = image.getpixel(dot)
            color_d=color_d_arr[0]
            if 0 < color_d <= 60:
                image.putpixel(dot, (0,0,0))
            if 60 < color_d <= 200:
                image.putpixel(dot, (128,128,128))
            if 200 < color_d <= 255:
                image.putpixel(dot, (255,255,255))
    image.save("trimap1111.jpg")
    return image

trimap chart

4. Through the original figure and trimap Figure combines the background color to generate the final result

def to_background(org, resize_trimap, id_image, color):
    """
        org： Original picture 
        resize_trimap：trimap
        id_image： New picture 
        color:  The background color 
    """
    scale = 1.0
    image = load_image(org, "RGB", scale, "box")
    trimap = load_image(resize_trimap, "GRAY", scale, "nearest")
    im = Image.open(org)
    # estimate alpha from image and trimap
    try:
        alpha = estimate_alpha_cf(image, trimap)
    except:
        print(" abnormal ")
        print(org)
        a = "-1"
        return a
    new_background = Image.new('RGB', im.size, color_dict[color])
    new_background.save("bj.png")
    # load new background
    new_background = load_image("bj.png", "RGB", scale, "box")
    # estimate foreground from image and alpha
    foreground, background = estimate_foreground_ml(image, alpha, return_background=True)
    # blend foreground with background and alpha
    new_image = blend(foreground, new_background, alpha)
    save_image(id_image, new_image)
    return id_image

Result chart

thus , The function of generating solid color ID photos has been realized .
Next use Vue Write a program interface to provide user interaction , Because the ID photo has a specific size , With the help of vue-croppe It allows users to cut the solid color certificate photo into one inch or two inches at the front end .

There are too many codes to show , You need to jump to the end of the text to get the source code .

The interface implementation effect is as follows ：

Source code is as follows ：
Back end ：https://github.com/QiuYang01/profilepicture-server
front end ：https://github.com/QiuYang01/profilepicture-client