-
Notifications
You must be signed in to change notification settings - Fork 1
/
frames2nparray.py
90 lines (70 loc) · 2.97 KB
/
frames2nparray.py
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
import os
import cv2
import numpy as np
from tqdm import tqdm
from PIL import Image
REBUILD_DATA = True
class LeftToRight():
IMG_SIZE = (1280,720) # Size of the movie frames
LEFT = "left/"
RIGHT = "right/"
training_data = []
size = 640, 256 # Size of NP arrays
box = (0,100,1280,620) # Cropping of movie frames.
def paths_in_npy_out(self, path_LM, path_LN, path_LO, path_R, box, size):
try:
LN = Image.open(path_LN).crop(box).resize(size, Image.ANTIALIAS)
RN = Image.open(path_R).crop(box).resize(size, Image.ANTIALIAS)
first = cv2.imread(path_LM, cv2.IMREAD_GRAYSCALE)
img = cv2.imread(path_LN, cv2.IMREAD_GRAYSCALE)
last = cv2.imread(path_LO, cv2.IMREAD_GRAYSCALE)
LM_flow = cv2.calcOpticalFlowFarneback(img, first, None, 0.5, 2, 15, 1, 5, 1.1, 0)[100:620, :, :]
LO_flow = cv2.calcOpticalFlowFarneback(img, last, None, 0.5, 2, 15, 1, 5, 1.1, 0)[100:620, :, :]
LM_flow = cv2.resize(LM_flow, dsize=(640, 256))
LO_flow = cv2.resize(LO_flow, dsize=(640, 256))
LN_flow = np.concatenate((LM_flow, np.array(LN), LO_flow, np.array(RN)), axis=2)
return LN_flow
except Exception as e:
print(str(e))
return None
def make_training_data(self, movie_details):
LEFT = self.LEFT
RIGHT = self.RIGHT
size = self.size
movie = movie_details[0]
box = self.box
# First build left 3 img dataset.
k = 0
j = 0
for i in tqdm(range(movie_details[1], movie_details[2])):
#Reads in files from directories$ holding 720p frames of left and right stereo images
path_LM = os.path.join(LEFT+movie, os.listdir(LEFT+movie)[i-1])
path_LN = os.path.join(LEFT+movie, os.listdir(LEFT+movie)[i])
path_LO = os.path.join(LEFT+movie, os.listdir(LEFT+movie)[i+1])
path_R = os.path.join(RIGHT+movie, os.listdir(RIGHT+movie)[i])
try:
self.training_data.append(self.paths_in_npy_out(path_LM, path_LN, path_LO, path_R, box, size))
except Exception as e:
print(str(e))
pass
j+=1
if j == 700 or i+1 == movie_details[2]:
k+=1
np.save(f"numpy_flow/{movie}_wflow-{k}.npy", self.training_data)
self.training_data.clear()
j = 0
if REBUILD_DATA:
frame_ends = [['prometh', 0, -1]]
lefttoright = LeftToRight()
for i in frame_ends:
lefttoright.make_training_data(i)
frame_ends = [['hobbit2', 1500, 108150],
['hobbit2p2', 0, 109000],
['hobbit3', 1500, 118350],
['hobbit3p2', 0, 70900],
['residentevilafterlife', 900, 129375],
['residentevilretribution', 1420, 123150],
['transformer_knight', 0, 210700],
['transformersmoon', 0, 210900],
['xmen_apocalypse', 1400, 193000]]
# %%