File size: 3,427 Bytes
21b679c efdedf4 21b679c efdedf4 65a9912 efdedf4 21b679c b12ade4 65a9912 b12ade4 21b679c b12ade4 21b679c | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 | '''
1. Accept image via Gradio
* image will come out an a numpy array
Done!
2. resize image
* Solved with PIL
3. split image into blocks
* Solved with PIL
4. find color of the square
* potentially solved with PIL
6. Find image with similar color profile
*need database
7. Place image in the new output image
* can be done with PIL
'''
import gradio as gr
import numpy as np
from PIL import Image
from SSIM_PIL import compare_ssim
# Define a function that applies a sepia filter to an image.
def mosiac(input_img):
pil_img = Image.fromarray(input_img)
width, height = pil_img.size
#resizing the image
if width != height:
if width < height:
#make the dimensions the width
dims = np.pow(np.floor(np.sqrt(np.sqrt(width))), 4)
elif height < width:
#make dimensions the height
dims = np.pow(np.floor(np.sqrt(np.sqrt(height))), 4)
dims = int(dims)
box = (dims, dims)
cropped_original = pil_img.resize(box)
pil_img = cropped_original.copy()
quant_img = pil_img.quantize(colors = 100)
#putting a grid on the image
#gridsize
grid_size = 16
quant_img = pil_img.quantize(colors = 100)
#putting a grid on the image
#gridsize
grid_size = 150
#size of each image
img_size = int(width/grid_size)
for column in range(0, grid_size):
for row in range(0, grid_size):
box = (int((row - 1) * img_size), int((column - 1) * img_size), int(row * img_size), int(column * img_size))
tile = quant_img.crop(box)
tile = tile.quantize(colors=3, method=2)
tile_rgb = tile.getpalette()[:9]
first_flag = True
with open("imageRGBs.txt", "r") as database:
for line in database:
rgb, img_num = line.split(';')
rgb = rgb.split(' ')
rgb = list(map(int, rgb))
difference = np.abs((tile_rgb[0] - rgb[0]) + (tile_rgb[1] - rgb[1]) + (tile_rgb[2] - rgb[2]))
difference = difference + np.abs((tile_rgb[3] - rgb[3]) + (tile_rgb[4] - rgb[4]) + (tile_rgb[5] - rgb[5])) * .5
difference = difference + np.abs((tile_rgb[6] - rgb[6]) + (tile_rgb[7] - rgb[7]) + (tile_rgb[8] - rgb[8])) * .333
if first_flag:
min_diff = difference
min_img_num = img_num.replace("\n", "")
first_flag = False
else:
if (min_diff > difference):
min_diff = difference
min_img_num = img_num.replace("\n", "")
replace_image = Image.open("images/" + min_img_num)
replace_image = replace_image.resize((img_size, img_size))
pil_img.paste(replace_image, box)
ssim_score = compare_ssim(pil_img, cropped_original)
print(ssim_score)
return pil_img
# Create a Gradio interface that takes an image input, runs it through the sepia function, and returns the filtered image.
demo = gr.Interface(mosiac, gr.Image(), "image")
# Launch the interface.
if __name__ == "__main__":
demo.launch(show_error=True) |