mirror/DeepCreamPy
1
0
Fork 0
mirror of https://github.com/Deepshift/DeepCreamPy.git synced 2025-02-23 09:40:59 +00:00
Code Issues Packages Projects Releases Wiki Activity

begin move of image operations to utils.py

Browse Source
This commit is contained in:
deeppomf 2018-11-08 09:12:19 -05:00
parent e886a79a6b
commit 13bf7e62c0
2 changed files with 15 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

20
decensor.py
View File

@ -6,7 +6,7 @@ from copy import deepcopy
import config
from libs.pconv_hybrid_model import PConvUnet
from libs.flood_fill import find_regions, expand_bounding
from libs.utils import *
class Decensor:
@ -81,8 +81,7 @@ class Decensor:
alpha_channel = np.expand_dims(alpha_channel, axis =-1)
ori = ori.convert('RGB')
ori_array = np.asarray(ori)
ori_array = np.array(ori_array / 255.0)
ori_array = image_to_array(ori)
ori_array = np.expand_dims(ori_array, axis = 0)
if self.is_mosaic:
@ -110,18 +109,16 @@ class Decensor:
mask_img = Image.fromarray(mask_reshaped.astype('uint8'))
#resize the cropped images
crop_img = crop_img.resize((512, 512))
crop_img_array = np.asarray(crop_img)
crop_img_array = crop_img_array / 255.0
crop_img_array = image_to_array(crop_img)
crop_img_array = np.expand_dims(crop_img_array, axis = 0)
#resize the mask images
mask_img = mask_img.crop(bounding_box)
mask_img = mask_img.resize((512, 512))
# mask_img.show()
#convert mask_img back to array
mask_array = np.asarray(mask_img)
mask_array = np.array(mask_array / 255.0)
mask_array = image_to_array(mask_img)
#the mask has been upscaled so there will be values not equal to 0 or 1
#mask_array[mask_array < 0.01] = 0
mask_array[mask_array > 0] = 1
mask_array = np.expand_dims(mask_array, axis = 0)
@ -142,12 +139,13 @@ class Decensor:
pred_img = Image.fromarray(pred_img_array.astype('uint8'))
# pred_img.show()
pred_img = pred_img.resize((bounding_width, bounding_height), resample = Image.BICUBIC)
pred_img_array = np.asarray(pred_img)
pred_img_array = pred_img_array / 255.0
pred_img_array = image_to_array(pred_img)
# print(pred_img_array.shape)
pred_img_array = np.expand_dims(pred_img_array, axis = 0)
# copy the decensored regions into the output image
for i in range(len(ori_array)):
for col in range(bounding_width):
for row in range(bounding_height):
@ -159,7 +157,7 @@ class Decensor:
output_img_array = output_img_array * 255.0
#restore the alpha channel
#restore the alpha channel if the image had one
if has_alpha:
#print(output_img_array.shape)
#print(alpha_channel.shape)

6
libs/flood_fill.py → libs/utils.py
View File

@ -1,5 +1,11 @@
import numpy as np
from PIL import Image, ImageDraw
#convert PIL image to numpy array
def image_to_array(image):
array = np.asarray(image)
return np.array(array / 255.0)
#find strongly connected components with the mask color
def find_regions(image):
pixel = image.load()