#Define image transformers
print('Shape mean_array : ' + str(mean_array.shape))
print('Shape net: ' + str(net.blobs[input_layer].data.shape))
net.blobs[input_layer].reshape(
1, # batch size
3, # channel
IMAGE_WIDTH,
IMAGE_HEIGHT) # image size
transformer = caffe.io.Transformer(
{input_layer: net.blobs[input_layer].data.shape})
transformer.set_mean(input_layer, mean_array)
transformer.set_transpose(input_layer, (2, 0, 1))
'''
Making predicitions
'''
im_files = imagelist_in_depth(r'c:\Users\szmike\Documents\DATA\studio2')
im_file = im_files[1]
image = cv2.imread(im_file)
print(im_file)
img = cv2.imread(im_file, cv2.IMREAD_COLOR)
img = transform_img(img, img_width=IMAGE_WIDTH, img_height=IMAGE_HEIGHT)
net.blobs[input_layer].data[...] = transformer.preprocess(input_layer, img)
out = net.forward()
print(out)
cv2.imshow("Output", image)
cv2.waitKey(0)
#Reading image paths
# Arrays to store object points and image points from all the images.
objpoints = [] # 3d point in real world space
imgpoints = [] # 2d points in image plane.
with np.load(calib_mtx_file) as X:
mtx, dist, _, _ = [X[i] for i in ('mtx','dist','rvecs','tvecs')]
fourcc = cv2.VideoWriter_fourcc('M', 'J', 'P', 'G')
out_file='calibration'+'.avi'
out = cv2.VideoWriter(out_file,fourcc, 1, th_size,True)
image_files=fh.imagelist_in_depth(image_dir,level=1,date_sort=False)
image_files=[image_file for image_file in image_files if os.path.basename(image_file).startswith('cam')]
gridsize=10
clahe = cv2.createCLAHE(clipLimit=2.0,tileGridSize=(gridsize,gridsize))
i=0
for i, image_file in enumerate(image_files):
if not i%1==0:
continue
print(image_file)
im=cv2.imread(image_file)
reader =csv.DictReader(open(typedict_3_file, 'rt'), delimiter=';')
for row in reader:
type_dict_3[row['type']]=row['label']
# LOAD MODEL
pred=load_model(model_file)
image_mean = 128
def keysWithValue(aDict, target):
return sorted(key for key, value in aDict.items() if target == value)
image_list_indir = imagelist_in_depth(image_dir,level=1)
preds=[]
for i, im_name in enumerate(image_list_indir):
# i=0
# im_name=image_list_indir[i]
image_file=os.path.join(image_dir,im_name)
img = Image.open(image_file).convert('RGB')
img_square=crop.crop(img)
im=np.asarray(img_square)
with warnings.catch_warnings():
warnings.simplefilter("ignore")
data = img_as_ubyte(resize(im, (imgSize,imgSize), order=1))
rgb_image=data.astype('float32')
# -*- coding: utf-8 -*-
"""
Created on Fri Mar 17 20:12:01 2017
@author: SzMike
"""
import file_helper
import os
import json
#base_folder = r'd:\DATA\Alinari\'
base_folder = os.path.curdir
user = '******' # picturio
image_dir = os.path.join(r'd:\DATA\RealEstate\117094')
image_list_file = os.path.join(base_folder, 'input', 'image_list.json')
image_list_indir = file_helper.imagelist_in_depth(image_dir, level=0)
image_label = {}
for image in image_list_indir:
image_label[image] = '0'
with open(image_list_file, 'w') as imagelistfile:
json.dump(image_list_indir, imagelistfile)
def new_list(self, image_dir):
self.current_dir = image_dir
self.image_list = file_helper.imagelist_in_depth(image_dir, level=0)
self.is_Features_ready = False
self.is_Scores_ready = False
from sklearn import cluster
import aes_Picturio
import aes_AADB
base_dir=os.path.join('e:','OneDrive','AES','Photo_DB')
image_dir=os.path.join(base_dir,'RealEstate')
save_dir=os.path.join(r'D:\DATA\RealEstate\AES')
score_file_merged=os.path.join(image_dir,'scores_merged.csv')
"""
Scoring model - manual scores
"""
image_list=fh.imagelist_in_depth(image_dir,level=1)
"""
Class names from folder names
"""
class_names=[os.path.dirname(f).split('\\')[-1] for f in image_list]
df_db = pd.DataFrame(data={'Filename':image_list,'Class name':class_names})
df_scores=df_db.copy()
"""
Scoring model - AADB
"""
scoring=aes_AADB.scoring()
image_all_scores=scoring.get_scores(df_db['Filename'].values)
#scores = [image_all_scores[i]['AestheticScore'] for i in range(len(image_all_scores)) if image_all_scores[i]['AestheticScore']!= 'None']