def main(img):
img1 = img
cloth_path = []
with open('clothes_match_labeled_data.txt',
encoding='gbk',
errors='ignore') as f:
for line in f:
name = line.split(':')[0]
path = 'my_training_shirts/' + name + '.jpg'
cloth_path.append(path)
cloth_path = cloth_path[1:]
similary_list = []
for img2 in cloth_path:
print('calculating the similarity with ' + img2)
sim1 = img_similarity(img1, img2)
# sim2 = compare_image(img1, img2)
sim3 = histsimilar.calc_similar_by_path(img1, img2)
similary = 0.4 * sim1 + 0.6 * sim3
similary_list.append(similary)
similary_dic = {}.fromkeys(cloth_path)
i = 0
for key, value in similary_dic.items():
similary_dic[key] = similary_list[i]
i += 1
best_img = max(similary_dic, key=similary_dic.get)
best_img = best_img.split('/')[-1]
best_img = best_img.split('.')[0]
return best_img
def pic_operate(case_name, is_update, size, drive):
name = case_name
pic_drive = drive
pic_size = copy.deepcopy(size)
pic_expect_path = path.get_multi_platform(PATH_PIC + "/pic/expect/%s.png"%name)
pic_fact_path = path.get_multi_platform(PATH_PIC + "/pic/fact/%s.png"%name)
result_path = path.get_multi_platform(PATH_PIC + "/result/%s.png"%name)
pic_size[2] = pic_size[2] + pic_size[0]
pic_size[3] = pic_size[3] + pic_size[1]
logger.info("截图起点坐标: " + str(pic_size[0]) + "," + str(pic_size[1]) + " 截图终点坐标: " + str(pic_size[2]) + "," + str(pic_size[3]))
global similar_rate
if is_update == "Y":
pic_drive.get_screenshot_as_file(pic_expect_path)
Image.open(pic_expect_path).crop(pic_size).save(pic_expect_path)
similar_rate = 100
elif is_update == "N":
pic_drive.get_screenshot_as_file(pic_fact_path)
Image.open(pic_fact_path).crop(pic_size).save(pic_fact_path)
pic0 = Image.open(pic_expect_path)
pic1 = Image.open(pic_fact_path)
pic_joint(pic0, pic1).save(result_path)
if is_reCompare == 0:
#上报图片所属用例
if result.case_pic.has_key(exe_status.exe_case_title):
result.case_pic[exe_status.exe_case_title] += name + ".png" + "|"
else:
result.case_pic[exe_status.exe_case_title] = name + ".png" + "|"
similar_rate = histsimilar.calc_similar_by_path(pic_fact_path,pic_expect_path)*100
return similar_rate
EPSILON = 1e-08
STYLE_SCALE = 1.0
ITERATIONS = 2101
initial_noiseblend = 1.0
VGG_PATH = 'imagenet-vgg-verydeep-19.mat'
POOLING = 'max'
img_cor = []
content = './images/ex5.jpg'
styles = []
#Get all the style as a array
for i in range(1, 18):
img_cor.append([
'./styles/%d.jpg' % i,
histsimilar.calc_similar_by_path(content, './styles/%d.jpg' % i)
])
img_cor.sort(key=lambda x: x[1], reverse=True)
#Take the possible related style
styles.append(img_cor[0])
print(img_cor[0][0])
styles.append(img_cor[1])
print(img_cor[1][0])
def main():
content_image = imread(content)
style_images = [imread(style[0]) for style in styles] #Accepted all style
style_blend_weights = [style[1] for style in styles]
target_shape = content_image.shape
cloth_path = []
with open('clothes_match_labeled_data.txt',
encoding='gbk',
errors='ignore') as f:
for line in f:
name = line.split(':')[0]
path = '/my_training_shirts/' + name + '.jpg'
cloth_path.append(path)
cloth_path = cloth_path[1:]
similary_list = []
for img2 in cloth_path:
print('calculating the similarity with ' + img2)
sim1 = img_similarity(img1, img2)
# sim2 = compare_image(img1, img2)
sim3 = histsimilar.calc_similar_by_path(img1, img2)
similary = 0.4 * sim1 + 0.6 * sim3
similary_list.append(similary)
similary_dic = {}.fromkeys(cloth_path)
i = 0
for key, value in similary_dic.items():
similary_dic[key] = similary_list[i]
i += 1
best_img = max(similary_dic, key=similary_dic.get)
print(best_img)
# cv2.namedWindow("new", cv2.WINDOW_NORMAL)
# cv2.imshow("new", img1)
# cv2.waitKey()
# img2_path = '4.jpg'
def similar():
#path = r'testpic/TEST%d/%d.JPG'
for i in xrange(1, 10):
print 'test_case_%d: %.3f%%'%(i, \
histsimilar.calc_similar_by_path('testpic/TEST%d/%d.JPG'%(i, 1), 'testpic/TEST%d/%d.JPG'%(i, 2))*100)
import os
import sys
import numpy as np
import scipy.io
import scipy.misc
import tensorflow as tf
import histsimilar
COR = []
OUTPUT_DIR = './outputs/'
STYLE_IMAGE = []
CONTENT_IMAGE = './images/ex1.jpg'
for i in range(1, 17):
COR.append([
'./styles/%d.jpg' % i,
histsimilar.calc_similar_by_path(CONTENT_IMAGE, './styles/%d.jpg' % i)
])
COR.sort(key=lambda x: x[1], reverse=True)
STYLE_IMAGE.append(COR[0][0])
print(COR[0][0])
STYLE_IMAGE.append(COR[1][0])
print(COR[1][0])
if COR[1][1] > 0.4:
STYLE_IMAGE.append(COR[2][0])
print(COR[2][0])
img = scipy.misc.imread(CONTENT_IMAGE).astype(np.float)
IMAGE_WIDTH = img.shape[1]
IMAGE_HEIGHT = img.shape[0]
