Log In Sign Up
Models:
Mathew Swanson
Mathew-S/
Master
0
Downloads: 1
[8eeb5a]: / utils / plotting.py

Download this file

96 lines (83 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
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
import matplotlib.pyplot as plt

import numpy as np

import pandas as pd

from scipy.stats import linregress, iqr





def get_best_epoch_from_id_split(df):

    mean_ious_per_epoch = [np.mean(df[df["epoch"] == i]["iou"]) for i in np.unique(df["epoch"])]

    return np.argmax(mean_ious_per_epoch)





def plot_std_iou_correlation(df, best=True):

    iou_std = []

    iou_avg = []

    for i in np.unique(df["id"]):

        id_df = df[df["id"] == i]

        id_best_epoch = get_best_epoch_from_id_split(id_df)

        if best:

            ious = id_df[id_df["epoch"] == id_best_epoch]["iou"]

            iou_std.append(np.std(ious))

            iou_avg.append(np.mean(ious))

        else:

            iou_std.append(np.nanstd(id_df[id_df["iou"] > 0.01]["iou"]))

            iou_avg.append(np.nanmean(id_df[id_df["iou"] > 0.01]["iou"]))



    plt.scatter(iou_std, iou_avg)

    print(iou_avg)

    result = linregress(iou_std, iou_avg)

    m, c, r, p, se = result

    print(result)

    x = np.linspace(np.min(iou_std), np.max(iou_std), 100)

    plt.plot(x, m * x + c, label="R = {}".format(r))

    plt.legend()

    plt.ylabel("Mean of IoUs")

    plt.xlabel("StDev of IoUs")

    plt.show()





def compare_variances(baseline_df, stresstest_df):

    baseline_mean_ious = []

    stresstest_mean_ious = []

    for id in np.unique(stresstest_df["id"]):

        baseline_ious = baseline_df[baseline_df["id"] == id]["iou"]

        stresstest_ious = stresstest_df[stresstest_df["id"] == id]["iou"]

        baseline_mean_ious.append(np.mean(baseline_ious))

        stresstest_mean_ious.append(np.mean(stresstest_ious))

    plt.hist(baseline_mean_ious, alpha=0.5, label=f"Baseline; iqr={iqr(baseline_mean_ious)}")

    plt.hist(stresstest_mean_ious, alpha=0.5, label=f"Stresstest; iqr={iqr(stresstest_mean_ious)}")

    plt.legend()

    plt.xlabel("Mean IoU")

    plt.ylabel("P(Mean IoU)")



    # plt.show()





if __name__ == '__main__':

    df = pd.read_csv("logs/ious.log")

    # df_untrained = pd.read_csv("logs/no-pretrain-ious.log")

    df_etis = pd.read_csv("logs/etis-ious.log")

    df_etis_stresstest = pd.read_csv("logs/etis-stresstest-results.log")

    df_kvasir_stresstest = pd.read_csv("logs/stresstest-results.log")

    df_kvasir = pd.read_csv("logs/kvasir-baseline.log")

    df_untrained = pd.read_csv("logs/kvasir-no-pretrain-baseline.log")

    # plot_std_iou_correlation(df, best=True)

    # plot_std_iou_correlation(df_untrained, best=True)

    # plot_std_iou_correlation(df_stresstest, best=False)

    # compare_variances(df_etis, df_etis_stresstest)

    # compare_variances(df, df_kvasir_stresstest)

    compare_variances(df_kvasir, df_untrained)

    plt.show()



    plot_std_iou_correlation(df_untrained, best=False)

    plot_std_iou_correlation(df, best=False)



    # compare_variances(df, df)

    # ious = df["iou"]

    # stds = []

    # mean_ious = []

    # for i in np.unique(df["id"]):

    #     plt.hist(ious[df["id"] == i][df["epoch"] == 175], alpha=0.5, bins=np.linspace(0, 1, 25), label=i, density=True)

    #     mean_ious.append(np.mean(ious[df["id"] == i][df["epoch"] == 175]))

    #     stds.append(np.std(ious[df["id"] == i][df["epoch"] == 175]))

    # # plt.legend()

    # plt.tight_layout()

    # plt.show()

    # print(stds)

    # plt.hist(stds)

    # plt.show()

    #

    # plt.scatter(stds, mean_ious)

    # m, c, r, p, se = linregress(stds, mean_ious)

    # print(m, c)

    # plt.xlabel("Standard deviation")

    # plt.ylabel("IoU")

    # plt.show()