-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmain.py
177 lines (151 loc) · 6.73 KB
/
main.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
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
import util
import const
import sift
import segmentation as seg
def filter_correspondences_and_boxes(correspondences, boxes_0, boxes_1):
filtered_correspondences = []
filtered_boxes_0 = []
filtered_boxes_1 = []
for correspondence in correspondences:
point_0 = correspondence[0]
point_1 = correspondence[1]
box_0_containing_point = util.get_box_containing_point(point_0, boxes_0)
box_1_containing_point = util.get_box_containing_point(point_1, boxes_1)
if (box_0_containing_point is not None and
box_1_containing_point is not None and
box_0_containing_point not in filtered_boxes_0 and
box_1_containing_point not in filtered_boxes_1
):
filtered_correspondences.append(correspondence)
filtered_boxes_0.append(box_0_containing_point)
filtered_boxes_1.append(box_1_containing_point)
return (filtered_correspondences, filtered_boxes_0, filtered_boxes_1)
def compare_two_images(img_0, img_1, num_sift_features, sift_correspondence_ratio):
img0_keypoints, img0_descriptors = sift.run_sift(img_0, num_features=num_sift_features)
img1_keypoints, img1_descriptors = sift.run_sift(img_1, num_features=num_sift_features)
img0_img1_correspondences = sift.find_sift_correspondences(img0_keypoints,
img0_descriptors,
img1_keypoints,
img1_descriptors,
ratio=sift_correspondence_ratio)
_, inliers, outliers = sift.ransac(
correspondences=img0_img1_correspondences,
num_iterations=50,
num_sampled_points=6,
threshold=3
)
img0_img1_correspondences = inliers
img0_contours = seg.find_contours(img_0)
img1_contours = seg.find_contours(img_1)
img0_boxes = seg.merge_overlapping_bounding_boxes(
bounding_boxes=seg.find_bounding_boxes(img0_contours, padding=0)
)
img1_boxes = seg.merge_overlapping_bounding_boxes(
bounding_boxes=seg.find_bounding_boxes(img1_contours, padding=0)
)
filtered_correspondences, filtered_boxes_0, filtered_boxes_1 = filter_correspondences_and_boxes(
correspondences=img0_img1_correspondences,
boxes_0=img0_boxes,
boxes_1=img1_boxes
)
return filtered_correspondences, filtered_boxes_0, filtered_boxes_1
def track_particle_motion(frames, num_sift_features, sift_correspondence_ratio, should_save_comparisons=False, should_save_ROIs=False, margins=None):
GLOBAL_MAP = {}
GLOBAL_LIST = []
for frame_n in range(len(frames)-1):
frame_n_filename = frames[frame_n]
frame_n_1_filename = frames[frame_n+1]
imgN = util.read_image(frame_n_filename)
imgN_1 = util.read_image(frame_n_1_filename)
if margins != None:
imgN = util.get_image_center(image=imgN, margins=margins)
imgN_1 = util.get_image_center(image=imgN_1, margins=margins)
corrs, boxes_N, boxes_N_1 = compare_two_images(
imgN,
imgN_1,
num_sift_features=num_sift_features,
sift_correspondence_ratio=sift_correspondence_ratio,
)
if should_save_comparisons:
imgN_with_boxes = seg.draw_bounding_boxes(imgN, bounding_boxes=boxes_N, border_color=const.COLOR_GREEN)
img_N_1_with_boxes = seg.draw_bounding_boxes(imgN_1, bounding_boxes=boxes_N_1, border_color=const.COLOR_GREEN)
sift.plot_correspondences(
image1=imgN_with_boxes,
image2=img_N_1_with_boxes,
correspondences=corrs,
color=const.COLOR_RED,
file_name=f'{util.get_no_extension_filename(frame_n_filename)}_to_{util.get_no_extension_filename(frame_n_1_filename)}'
)
# Initialize the GLOBAL objects with frame_0 information
if frame_n == 0:
for box_0 in boxes_N:
list_index = len(GLOBAL_LIST)
GLOBAL_MAP[box_0] = list_index
ROI_area = seg.get_non_zero_pixel_area(
image=seg.crop_image(image=imgN, box=box_0)
)
GLOBAL_LIST.append([])
GLOBAL_LIST[list_index].append((ROI_area, (0, box_0)))
print(f'\nGLOBAL_MAP: {GLOBAL_MAP}')
curr_map = {} #object to collect current maaping
for corr, box_N, box_N_1 in zip(corrs, boxes_N, boxes_N_1):
imgN_ROI = seg.crop_image(imgN, box_N)
imgN_1_ROI = seg.crop_image(imgN_1, box_N_1)
imgN_ROI_area = seg.get_non_zero_pixel_area(imgN_ROI)
imgN_1_ROI_area = seg.get_non_zero_pixel_area(imgN_1_ROI)
print(f'({imgN_ROI_area},{imgN_1_ROI_area}) img{frame_n}_{box_N} to img{frame_n+1}_{box_N_1} via {util.get_readable_correspondence(corr)}')
if should_save_ROIs: util.write_image(imgN_ROI, f'img{frame_n}_{box_N}')
list_index = -1
if box_N in GLOBAL_MAP:
list_index = GLOBAL_MAP[box_N]
else:
list_index = len(GLOBAL_LIST)
GLOBAL_LIST.append([])
GLOBAL_LIST[list_index].append((imgN_1_ROI_area, (frame_n+1, box_N_1)))
curr_map[box_N_1] = list_index
GLOBAL_MAP = curr_map
GLOBAL_LIST = list(reversed(sorted(GLOBAL_LIST, key=len)))
print(GLOBAL_LIST)
return GLOBAL_LIST
def bulk_carbon(frames, margins=None, should_save_images=False):
frame_to_carbon_map = {}
for frame in frames:
frame_image = util.read_image(frame)
if margins != None:
frame_image = util.get_image_center(image=frame_image, margins=margins)
contours = seg.find_contours(image=frame_image)
boxes = seg.merge_overlapping_bounding_boxes(
bounding_boxes=seg.find_bounding_boxes(contours=contours, padding=0)
)
if should_save_images:
util.write_image(
image=seg.draw_bounding_boxes(image=frame_image, bounding_boxes=boxes, border_color=const.COLOR_GREEN),
file_output_name=f'{util.get_no_extension_filename(frame)}_boxes'
)
total_pixel_area_for_frame = 0
num_particles = len(boxes)
for box in boxes:
pixel_area = seg.get_non_zero_pixel_area(
image=seg.crop_image(image=frame_image, box=box)
)
total_pixel_area_for_frame += pixel_area
print(f'Frame: {frame} has total particle pixel area: {total_pixel_area_for_frame} and total # of particles: {num_particles}')
frame_to_carbon_map[frame] = (total_pixel_area_for_frame, num_particles)
print(frame_to_carbon_map)
return frame_to_carbon_map
if __name__ == "__main__":
MARGINS_TO_USE = (300, 275, 400, 175)
FRAMES_TO_ANALYZE = const.MINION_3_FRAMES[:-1]
track_particle_motion(
frames=FRAMES_TO_ANALYZE,
num_sift_features=3000,
sift_correspondence_ratio=0.6,
should_save_comparisons=False,
should_save_ROIs=False,
margins=MARGINS_TO_USE
)
bulk_carbon(
frames=FRAMES_TO_ANALYZE,
margins=MARGINS_TO_USE,
should_save_images=False,
)