Models:
Samuel Callahan
SCallahan/
lung-ultrasound-scan-targ
0
Downloads: 1
[c1b1c5]: / src / evaluation / compute_gt_position_diff.py

Download this file

71 lines (57 with data), 3.5 kB 

 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
# Compute the ground-truth difference between three different methods:



import os

import pickle

import numpy as np



# three methods:

cam1_cam2_diff = {'target1': {'x': [], 'y': [], 'z': []}, 'target2': {'x': [], 'y': [], 'z': []},

                  'target4': {'x': [], 'y': [], 'z': []}}

two_cam_cam1_diff = {'target1': {'x': [], 'y': [], 'z': []}, 'target2': {'x': [], 'y': [], 'z': []},

                     'target4': {'x': [], 'y': [], 'z': []}}

two_cam_cam2_diff = {'target1': {'x': [], 'y': [], 'z': []}, 'target2': {'x': [], 'y': [], 'z': []},

                     'target4': {'x': [], 'y': [], 'z': []}}



data_path = '../data'



# loop through the data folder

for SUBJECT_NAME in os.listdir(data_path):

    subject_folder_path = os.path.join(data_path, SUBJECT_NAME)

    if os.path.isfile(subject_folder_path):

        continue



    # frontal points: target 1 & target 2

    front_folder_path = os.path.join(subject_folder_path, 'front')

    with open(front_folder_path + '/cam_1_gt.pickle', 'rb') as f:

        cam_1_gt = pickle.load(f)

    with open(front_folder_path + '/cam_2_gt.pickle', 'rb') as f:

        cam_2_gt = pickle.load(f)

    with open(front_folder_path + '/two_cam_gt.pickle', 'rb') as f:

        two_cam_gt = pickle.load(f)



    cam_1_tar1_gt, cam_1_tar2_gt = cam_1_gt['target_1'].flatten(), cam_1_gt['target_2'].flatten()

    cam_2_tar1_gt, cam_2_tar2_gt = cam_2_gt['target_1'].flatten(), cam_2_gt['target_2'].flatten()

    two_cam_tar1_gt, two_cam_tar2_gt = two_cam_gt['target_1'].flatten(), two_cam_gt['target_2'].flatten()



    # side point: target 4

    side_folder_path = os.path.join(subject_folder_path, 'side')

    with open(side_folder_path + '/cam_1_gt.pickle', 'rb') as f:

        cam_1_gt = pickle.load(f)

    with open(side_folder_path + '/cam_2_gt.pickle', 'rb') as f:

        cam_2_gt = pickle.load(f)

    with open(side_folder_path + '/two_cam_gt.pickle', 'rb') as f:

        two_cam_gt = pickle.load(f)



    cam_1_tar4_gt = cam_1_gt['target_4'].flatten()

    cam_2_tar4_gt = cam_2_gt['target_4'].flatten()

    two_cam_tar4_gt = two_cam_gt['target_4'].flatten()



    for target, cam1_tar, cam2_tar, two_cam_tar in zip(cam1_cam2_diff.keys(),

                                                       [cam_1_tar1_gt, cam_1_tar2_gt, cam_1_tar4_gt],

                                                       [cam_2_tar1_gt, cam_2_tar2_gt, cam_2_tar4_gt],

                                                       [two_cam_tar1_gt, two_cam_tar2_gt, two_cam_tar4_gt]):

        for i, axis in enumerate(cam1_cam2_diff[target].keys()):

            cam1_cam2_diff[target][axis].append(cam1_tar[i] - cam2_tar[i])

            two_cam_cam1_diff[target][axis].append(two_cam_tar[i] - cam1_tar[i])

            two_cam_cam2_diff[target][axis].append(two_cam_tar[i] - cam2_tar[i])



gt_diff_dict = {'cam1_cam2_diff': cam1_cam2_diff, 'two_cam_cam1_diff': two_cam_cam1_diff, 'two_cam_cam2_diff': two_cam_cam2_diff}

# with open('gt_diff.pickle', 'wb') as f:

#     pickle.dump(gt_diff_dict, f)



# Compute difference mean and std

for target in cam1_cam2_diff.keys():

    print("========================================================================")

    for gt_diff_name, gt_diff in zip(gt_diff_dict.keys(), gt_diff_dict.values()):

        for axis in cam1_cam2_diff[target].keys():

            print(f"{target}, {gt_diff_name}, {axis}-axis: mean = {np.mean(gt_diff[target][axis]) * 1000: .1f}")

            print(f"{target}, {gt_diff_name}, {axis}-axis: std = {np.std(gt_diff[target][axis]) * 1000: .1f}")