Object-Detection/performance_graphs.py at main · im-Amrith/Object-Detection · 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
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
import matplotlib.pyplot as plt
import numpy as np
import os

# Create output directory for graphs
os.makedirs('graphs', exist_ok=True)

# Set style for all plots
plt.style.use('ggplot')
plt.rcParams['figure.figsize'] = (10, 6)
plt.rcParams['font.size'] = 12

# 1. Detection accuracy vs. distance
distances = np.array([10, 20, 30, 40, 50, 60, 70, 80, 90, 100])
vehicle_accuracy = np.array([96.8, 95.3, 94.1, 92.7, 90.2, 86.5, 83.1, 80.4, 77.2, 75.0])
person_accuracy = np.array([97.2, 96.5, 94.8, 92.9, 89.6, 85.2, 81.7, 78.3, 74.5, 71.2])
traffic_light_accuracy = np.array([98.1, 97.3, 95.9, 93.8, 91.7, 88.2, 84.6, 80.1, 75.3, 72.0])

plt.figure()
plt.plot(distances, vehicle_accuracy, 'b-', linewidth=2, marker='o', label='Vehicles')
plt.plot(distances, person_accuracy, 'g-', linewidth=2, marker='s', label='Pedestrians')
plt.plot(distances, traffic_light_accuracy, 'r-', linewidth=2, marker='^', label='Traffic Lights')
plt.axhline(y=90, color='gray', linestyle='--', alpha=0.7)
plt.xlabel('Distance (meters)')
plt.ylabel('Detection Accuracy (%)')
plt.title('Detection Accuracy vs. Distance')
plt.grid(True, alpha=0.3)
plt.legend()
plt.tight_layout()
plt.savefig('graphs/accuracy_vs_distance.png', dpi=300)

# 2. Processing speed vs. number of objects
objects = np.array([5, 10, 15, 20, 25, 30, 35, 40])
fps = np.array([29.8, 27.2, 24.7, 21.3, 18.6, 16.2, 14.5, 12.8])

plt.figure()
plt.plot(objects, fps, 'b-', linewidth=2, marker='o')
plt.axhline(y=20, color='r', linestyle='--', alpha=0.7, label='Real-time threshold (20 FPS)')
plt.xlabel('Number of Objects per Frame')
plt.ylabel('Processing Speed (FPS)')
plt.title('Processing Speed vs. Number of Objects')
plt.grid(True, alpha=0.3)
plt.legend()
plt.tight_layout()
plt.savefig('graphs/speed_vs_objects.png', dpi=300)

# 3. Accuracy vs. lighting conditions
categories = ['Vehicles', 'Pedestrians', 'Traffic Lights', 'Traffic Light State']
daytime = np.array([94.2, 93.5, 96.1, 91.8])
nighttime = np.array([86.5, 85.2, 87.3, 82.1])
overcast = np.array([90.3, 89.7, 92.5, 86.4])

x = np.arange(len(categories))
width = 0.25

fig, ax = plt.subplots()
day_bars = ax.bar(x - width, daytime, width, label='Daytime', color='gold')
night_bars = ax.bar(x, nighttime, width, label='Nighttime', color='navy')
overcast_bars = ax.bar(x + width, overcast, width, label='Overcast', color='gray')

ax.set_xlabel('Detection Category')
ax.set_ylabel('Accuracy (%)')
ax.set_title('Detection Accuracy vs. Lighting Conditions')
ax.set_xticks(x)
ax.set_xticklabels(categories)
ax.legend()

# Add value labels on top of bars
def add_labels(bars):
    for bar in bars:
        height = bar.get_height()
        ax.annotate(f'{height:.1f}',
                    xy=(bar.get_x() + bar.get_width() / 2, height),
                    xytext=(0, 3),
                    textcoords="offset points",
                    ha='center', va='bottom', fontsize=8)

add_labels(day_bars)
add_labels(night_bars)
add_labels(overcast_bars)

plt.tight_layout()
plt.savefig('graphs/accuracy_vs_lighting.png', dpi=300)

# 4. Traffic light state recognition vs. distance
distances = np.array([10, 20, 30, 40, 50, 60, 70])
red_accuracy = np.array([97.2, 95.8, 94.5, 91.2, 87.4, 83.6, 79.8])
yellow_accuracy = np.array([94.1, 92.3, 89.7, 85.2, 80.8, 76.1, 71.5])
green_accuracy = np.array([95.6, 93.4, 91.3, 87.5, 82.9, 78.2, 73.8])
overall_accuracy = np.array([95.7, 93.8, 91.8, 88.0, 83.7, 79.3, 75.0])

plt.figure()
plt.plot(distances, red_accuracy, 'r-', linewidth=2, marker='o', label='Red Light')
plt.plot(distances, yellow_accuracy, 'y-', linewidth=2, marker='s', label='Yellow Light')
plt.plot(distances, green_accuracy, 'g-', linewidth=2, marker='^', label='Green Light')
plt.plot(distances, overall_accuracy, 'b-', linewidth=2, marker='x', label='Overall')
plt.axhline(y=90, color='gray', linestyle='--', alpha=0.7)
plt.xlabel('Distance (meters)')
plt.ylabel('Recognition Accuracy (%)')
plt.title('Traffic Light State Recognition Accuracy vs. Distance')
plt.grid(True, alpha=0.3)
plt.legend()
plt.tight_layout()
plt.savefig('graphs/traffic_light_accuracy_vs_distance.png', dpi=300)

# 5. Confusion Matrix Visualization for Traffic Light State
import matplotlib.pyplot as plt
import numpy as np
from matplotlib import colors

# Create a confusion matrix
cm = np.array([
    [312, 14, 3],
    [21, 187, 9],
    [5, 18, 173]
])

# Normalize the confusion matrix
cm_norm = cm / cm.sum(axis=1)[:, np.newaxis]

# Create the plot
fig, ax = plt.subplots(figsize=(8, 7))
im = ax.imshow(cm_norm, interpolation='nearest', cmap=plt.cm.Blues)

# Add colorbar
cbar = ax.figure.colorbar(im, ax=ax)
cbar.ax.set_ylabel('Normalized Frequency', rotation=-90, va="bottom")

# Add labels
classes = ['Red', 'Yellow', 'Green']
tick_marks = np.arange(len(classes))
ax.set_xticks(tick_marks)
ax.set_yticks(tick_marks)
ax.set_xticklabels(classes)
ax.set_yticklabels(classes)

# Add title and axes labels
ax.set_title('Traffic Light State Recognition Confusion Matrix')
ax.set_ylabel('Actual Class')
ax.set_xlabel('Predicted Class')

# Add text annotations
thresh = cm_norm.max() / 2.
for i in range(cm.shape[0]):
    for j in range(cm.shape[1]):
        text = f"{cm[i, j]}\n({cm_norm[i, j]:.2f})"
        ax.text(j, i, text,
                ha="center", va="center",
                color="white" if cm_norm[i, j] > thresh else "black")

fig.tight_layout()
plt.savefig('graphs/traffic_light_confusion_matrix.png', dpi=300)

print("All performance graphs generated successfully in the 'graphs' directory.")