-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathmodel-comparison.py
More file actions
154 lines (128 loc) · 8.5 KB
/
Copy pathmodel-comparison.py
File metadata and controls
154 lines (128 loc) · 8.5 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
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
import matplotlib.pyplot as plt
import numpy as np
from adjustText import adjust_text
plt.rcParams["font.family"] = "Serif"
plt.figure(figsize=(16, 8))
data1 = [
# model-name, year, OvAc, color, sym, type, offsetX, offsetY
# ["3DShapeNets", 2015+ 7/12, 77.0, '#f55', 'o', 'Convolutional Network', 0.0, 0.0],
["MVCNN", 2015+ 9/12, 90.1, '#f55', 'o', '_Convolutional Network', 0.0, 0.0],
["PointNet", 2017+ 4/12, 89.2, '#55f', '^', 'Multi-Layer Perceptron', 0.0, 0.0],
["PointNet++", 2017+ 6/12, 90.7, '#55f', '^', '_Multi-Layer Perceptron', 0.0, 0.0],
["PointCNN", 2018+11/12, 92.2, '#f55', 'o', '_Convolutional Network', 0.0, 0.0],
["PointConv", 2020+11/12, 92.5, '#f55', 'o', '_Convolutional Network', 0.0, 0.0],
["GeomGCNN", 2021+ 3/12, 95.9, '#5f5', 's', 'Graph Conv. Network', 0.0, 0.0],
["PointTrans.", 2021+ 9/12, 93.7, '#5ff', 'D', 'Transformer', -1.9, 0.0],
["PointViewGCN", 2021+ 9/12, 95.4, '#5f5', 's', '_Graph Conv. Network', 0.0, 0.0],
["PointBERT", 2022+ 6/12, 93.8, '#5ff', 'D', '_Transformer', 0.0, -0.3],
["PointMLP", 2022+11/12, 94.5, '#55f', '^', '_Multi-Layer Perceptron',-1.6, 0.0],
["PointNext", 2022+10/12, 94.0, '#55f', '^', '_Multi-Layer Perceptron', 0.0, 0.0],
["ReCon", 2023+ 5/12, 94.7, '#5ff', 'D', '_Transformer', 0.0, -0.1],
["OmniVec", 2023+11/12, 96.6, '#5ff', 'D', '_Transformer', 0.0, 0.0],
["ReCon++", 2024+ 3/12, 95.0, '#5ff', 'D', '_Transformer', 0.0, 0.0]
]
plt.subplot(2, 2, 1)
plt.title("Classification on ModelNet40", fontsize=15, fontweight='normal')
plt.xlabel("Year", labelpad=10, loc="center", fontsize=12, fontweight='normal')
plt.ylabel("OA (%)", labelpad=10, loc="center", fontsize=12, fontweight='normal')
plt.xlim(2015, 2026)
plt.ylim(88, 97)
plt.grid(True)
# labels = []
for entry in data1:
plt.plot(entry[1], entry[2], label=entry[5], color=entry[3], marker=entry[4], markersize=7)
plt.annotate(entry[0], (entry[1]+entry[6]+0.13, entry[2]+entry[7]-0.16), fontsize=10)
####
data2 = [
# model-name , year, OvAc, color, sym, type, offsetX, offsetY
["3DShapeNets", 2015+ 7/12, 00.00, '#f55', 'o', 'Convolutional Network', 0.0, 0.0],
["MVCNN", 2015+ 9/12, 00.00, '#f55', 'o', '_Convolutional Network', 0.0, 0.0],
["PointNet", 2017+ 4/12, 68.20, '#55f', '^', 'Multi-Layer Perceptron', 0.0, 0.0],
["PointNet++", 2017+ 6/12, 77.90, '#55f', '^', '_Multi-Layer Perceptron', 0.0, -0.6],
["PointCNN", 2018+11/12, 78.50, '#f55', 'o', '_Convolutional Network', 0.0, 0.0],
["PointConv", 2020+11/12, 00.00, '#f55', 'o', '_Convolutional Network', 0.0, 0.0],
["GeomGCNN", 2021+ 3/12, 00.00, '#5f5', 's', 'Graph Conv. Network', 0.0, 0.0],
["PointTrans.", 2021+ 9/12, 00.00, '#5ff', 'D', 'Transformer', 0.0, 0.0],
["PointViewGCN", 2021+ 9/12, 85.50, '#5f5', 's', '_Graph Conv. Network', -1.9, 0.0],
["PointBERT", 2022+ 6/12, 83.07, '#5ff', 'D', '_Transformer', 0.0, 0.0],
["PointMLP", 2022+11/12, 85.70, '#55f', '^', '_Multi-Layer Perceptron', 0.0, 0.0],
["PointNext", 2022+10/12, 87.70, '#55f', '^', '_Multi-Layer Perceptron', 0.0, 0.0],
["ReCon", 2023+ 5/12, 91.26, '#5ff', 'D', '_Transformer', 0.0, 0.0],
["OmniVec", 2023+11/12, 96.10, '#5ff', 'D', '_Transformer', 0.0, 0.6],
["ReCon++", 2024+ 3/12, 95.25, '#5ff', 'D', '_Transformer', 0.0, 0.0]
]
plt.subplot(2, 2, 2)
plt.title("Classification on ScanObjectNN", fontsize=15, fontweight='normal')
plt.xlabel("Year", labelpad=10, loc="center", fontsize=12, fontweight='normal')
plt.ylabel("OA (%)", labelpad=10, loc="center", fontsize=12, fontweight='normal')
plt.xlim(2017, 2026)
plt.ylim(67, 98)
plt.grid(True)
for entry in data2:
plt.plot(entry[1], entry[2], label=entry[5], color=entry[3], marker=entry[4], markersize=7)
plt.annotate(entry[0], (entry[1]+entry[6]+0.13, entry[2]+entry[7]-0.5), fontsize=10)
####
data3 = [
# model-name , year, mIoU, color, sym, type, offsetX, offsetY
["3DShapeNets", 2015+ 7/12, 00.00, '#f55', 'o', 'Convolutional Network', 0.0, 0.0],
["MVCNN", 2015+ 9/12, 00.00, '#f55', 'o', '_Convolutional Network', 0.0, 0.0],
["PointNet", 2017+ 4/12, 83.70, '#55f', '^', 'Multi-Layer Perceptron', 0.0, 0.0],
["PointNet++", 2017+ 6/12, 85.10, '#55f', '^', '_Multi-Layer Perceptron', 0.0, 0.0],
["PointCNN", 2018+11/12, 86.10, '#f55', 'o', '_Convolutional Network', 0.0, 0.0],
["PointConv", 2020+11/12, 85.70, '#f55', 'o', '_Convolutional Network', 0.0, 0.0],
["GeomGCNN", 2021+ 3/12, 89.10, '#5f5', 's', 'Graph Conv. Network', 0.0, 0.0],
["PointTrans.", 2021+ 9/12, 86.60, '#5ff', 'D', 'Transformer', 0.0, 0.0],
["PointViewGCN", 2021+ 9/12, 00.00, '#5f5', 's', '_Graph Conv. Network', 0.0, 0.0],
["PointBERT", 2022+ 6/12, 85.60, '#5ff', 'D', '_Transformer', 0.0, 0.0],
["PointMLP", 2022+11/12, 86.10, '#55f', '^', '_Multi-Layer Perceptron', 0.0, -0.1],
["PointNext", 2022+10/12, 87.10, '#55f', '^', '_Multi-Layer Perceptron', 0.0, 0.0],
["ReCon", 2023+ 5/12, 86.40, '#5ff', 'D', '_Transformer', 0.0, 0.0],
["OmniVec", 2023+11/12, 00.00, '#5ff', 'D', '_Transformer', 0.0, 0.0],
["ReCon++", 2024+ 3/12, 00.00, '#5ff', 'D', '_Transformer', 0.0, 0.0]
]
plt.subplot(2, 2, 3)
plt.title("Part Segmentation on ShapeNet-Part", fontsize=15, fontweight='normal', pad=10)
plt.xlabel("Year", labelpad=10, loc="center", fontsize=12, fontweight='normal')
plt.ylabel("mIoU (%)", labelpad=10, loc="center", fontsize=12, fontweight='normal')
plt.xlim(2017, 2026)
plt.ylim(83, 90)
plt.grid(True)
for entry in data3:
plt.plot(entry[1], entry[2], label=entry[5], color=entry[3], marker=entry[4], markersize=7)
plt.annotate(entry[0], (entry[1]+entry[6]+0.13, entry[2]+entry[7]-0.15), fontsize=10)
####
data4 = [
# model-name , year, mIoU, Color, Sym, type, offsetX, offsetY
["3DShapeNets", 2015+ 7/12, 00.00, '#f55', 'o', 'Convolutional Network', 0.0, 0.0],
["MVCNN", 2015+ 9/12, 00.00, '#f55', 'o', '_Convolutional Network', 0.0, 0.0],
["PointNet", 2017+ 4/12, 47.60, '#55f', '^', 'Multi-Layer Perceptron', 0.0, 0.0],
["PointNet++", 2017+ 6/12, 54.50, '#55f', '^', '_Multi-Layer Perceptron', 0.0, 0.0],
["PointCNN", 2018+11/12, 65.40, '#f55', 'o', '_Convolutional Network', 0.0, 0.0],
["DGCNN", 2019+ 6/12, 56.10, '#5f5', 's', '_Graph Conv. Network', 0.0, 0.0],
["PointConv", 2020+11/12, 00.00, '#f55', 'o', '_Convolutional Network', 0.0, 0.0],
["GeomGCNN", 2021+ 3/12, 70.10, '#5f5', 's', 'Graph Conv. Network', 0.0, 0.0],
["PointTrans.", 2021+ 9/12, 73.50, '#5ff', 'D', 'Transformer', 0.0, -0.5],
["PointViewGCN", 2021+ 9/12, 00.00, '#5f5', 's', '_Graph Conv. Network', 0.0, 0.0],
["PointBERT", 2022+ 6/12, 00.00, '#5ff', 'D', '_Transformer', 0.0, 0.0],
["PointMLP", 2022+11/12, 00.00, '#55f', '^', '_Multi-Layer Perceptron', 0.0, 0.0],
["PointNext", 2022+10/12, 74.90, '#55f', '^', '_Multi-Layer Perceptron', 0.0, -0.5],
["ReCon", 2023+ 5/12, 00.00, '#5ff', 'D', '_Transformer', 0.0, 0.0],
["OmniVec", 2023+11/12, 75.90, '#5ff', 'D', '_Transformer', 0.0, 0.0],
["ReCon++", 2024+ 3/12, 00.00, '#5ff', 'D', '_Transformer', 0.0, 0.0],
["Sonata + PTV3",2025+ 3/12, 82.30, '#5ff', 'D', '_Transformer', -2.0, 0.0]
]
plt.subplot(2, 2, 4)
plt.title("Semantic Segmentation on S3DIS", fontsize=15, fontweight='normal', pad=10)
plt.xlabel("Year", labelpad=10, loc="center", fontsize=12, fontweight='normal')
plt.ylabel("mIoU (%)", labelpad=10, loc="center", fontsize=12, fontweight='normal')
plt.xlim(2017, 2026)
plt.ylim(46, 84)
plt.grid(True)
for entry in data4:
plt.plot(entry[1], entry[2], label=entry[5], color=entry[3], marker=entry[4], markersize=7)
plt.annotate(entry[0], (entry[1]+entry[6]+0.13, entry[2]+entry[7]-0.7), fontsize=10)
####
plt.legend(loc='lower center', bbox_to_anchor=(-0.1, -0.5), title="Backbone Network Structures",
labelspacing=1, shadow=False, ncol=4, fontsize=10, title_fontsize=12)
plt.subplots_adjust(hspace=0.4, wspace=0.2)
plt.savefig("my_plot.pdf", format="pdf", bbox_inches="tight")