LandType-Sentinel2-Segmentation/utils.py at main · MohamedMostafa259/LandType-Sentinel2-Segmentation · GitHub

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
import numpy as np
import matplotlib.pyplot as plt
import tensorflow as tf
from IPython.display import clear_output
import itertools
import math


##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- #####


class PlotDiagnostics(tf.keras.callbacks.Callback):
  def on_train_begin(self, logs={}):
    self.i = 0
    self.x = []

    self.losses = []
    self.val_losses = []

    self.jaccard_coef = []
    self.val_jaccard_coef = []

    self.fig = plt.figure()
    self.logs = []

  def on_epoch_end(self, epoch, logs={}):
    self.logs.append(logs)
    self.x.append(self.i)

    self.losses.append(logs.get('loss'))
    self.val_losses.append(logs.get('val_loss'))

    self.jaccard_coef.append(logs.get('jaccard_coef'))
    self.val_jaccard_coef.append(logs.get('val_jaccard_coef'))

    self.i += 1

    plt.figure(figsize=(14,8))
    f, (graph1, graph2) = plt.subplots(1,2, sharex=True)

    clear_output(wait=True)

    graph1.set_yscale('log')
    graph1.plot(self.x, self.losses, label="loss")
    graph1.plot(self.x, self.val_losses, label="val_loss")
    graph1.legend()

    graph2.set_yscale('log')
    graph2.plot(self.x, self.jaccard_coef, label="jaccard_coef")
    graph2.plot(self.x, self.val_jaccard_coef, label="val_jaccard_coef")
    graph2.legend()

    plt.show()


##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- #####


def show_final_history(history):
    """Automatically plots all training/validation metrics from a Keras history object, two plots per row."""

    history_dict = history.history
    metrics = sorted(set(k.replace('val_', '') for k in history_dict if not k.startswith('val_')))
    num_metrics = len(metrics)

    # Compute number of rows needed if 2 plots per row
    num_cols = 2
    num_rows = math.ceil(num_metrics / num_cols)

    plt.style.use("ggplot")
    fig, axes = plt.subplots(num_rows, num_cols, figsize=(7 * num_cols, 5 * num_rows))
    axes = np.array(axes).reshape(-1)  # flatten in case it's 2D

    for idx, metric in enumerate(metrics):
        ax = axes[idx]
        ax.plot(history_dict[metric], 'r-', label=f'Training {metric}')
        val_key = f'val_{metric}'
        if val_key in history_dict:
            ax.plot(history_dict[val_key], 'g-', label=f'Validation {metric}')

        ax.set_title(metric.capitalize())
        ax.set_xlabel('Epochs')
        ax.set_ylabel(metric)
        ax.legend(loc='best')

    # Hide unused subplots if num_metrics is odd
    for j in range(len(metrics), len(axes)):
        fig.delaxes(axes[j])

    plt.tight_layout()
    plt.show()


##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- #####


# src: https://www.kaggle.com/code/apollo2506/eurosat-allbands-classification
def plot_learning_rate(history):
    plt.style.use("ggplot")
    plt.plot(np.arange(0, len(history.history['learning_rate'])), history.history['learning_rate'])
    plt.xlabel("Epochs")
    plt.ylabel("Learning Rate")
    plt.show()


##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- #####


def plot_confusion_matrix(cm, classes, title='Confusion Matrix', cmap=plt.cm.Blues):

    cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]
    plt.figure(figsize=(10, 10))
    plt.imshow(cm, interpolation='nearest', cmap=cmap)
    plt.title(title)
    plt.colorbar()

    tick_marks = np.arange(len(classes))
    plt.xticks(tick_marks, classes, rotation=90)
    plt.yticks(tick_marks, classes)

    fmt = '.2f'
    thresh = cm.max()/2.0

    for i,j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):

        plt.text(j,i, format(cm[i,j], fmt),
                horizontalalignment = 'center',
                color = "white" if cm[i,j] > thresh else "black")

    plt.ylabel("True Label")
    plt.xlabel("Predicted Label")
    plt.grid(False)


##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- ##### ----- #####