Spaces:
Running
on
CPU Upgrade
Running
on
CPU Upgrade
mischeiwiller
commited on
Commit
•
9526595
1
Parent(s):
3b00c47
fix: resolve IndexError in line fitting visualization
Browse files
app.py
CHANGED
@@ -1,13 +1,10 @@
|
|
1 |
-
# Example showing how to fit a 2d line with kornia / pytorch
|
2 |
import matplotlib.pyplot as plt
|
3 |
import torch
|
4 |
import matplotlib
|
5 |
matplotlib.use('Agg')
|
6 |
-
import matplotlib.pyplot as plt
|
7 |
import gradio as gr
|
8 |
from kornia.geometry.line import ParametrizedLine, fit_line
|
9 |
|
10 |
-
|
11 |
def inference(point1, point2, point3, point4):
|
12 |
std = 1.2 # standard deviation for the points
|
13 |
num_points = 50 # total number of points
|
@@ -16,7 +13,7 @@ def inference(point1, point2, point3, point4):
|
|
16 |
p0 = torch.tensor([point1, point2], dtype=torch.float32)
|
17 |
p1 = torch.tensor([point3, point4], dtype=torch.float32)
|
18 |
l1 = ParametrizedLine.through(p0, p1)
|
19 |
-
|
20 |
# sample some points and weights
|
21 |
pts, w = [], []
|
22 |
for t in torch.linspace(-10, 10, num_points):
|
@@ -25,37 +22,41 @@ def inference(point1, point2, point3, point4):
|
|
25 |
p2 += p2_noise
|
26 |
pts.append(p2)
|
27 |
w.append(1 - p2_noise.mean())
|
|
|
28 |
pts = torch.stack(pts)
|
29 |
w = torch.stack(w)
|
|
|
30 |
if len(pts.shape) == 2:
|
31 |
-
pts = pts
|
32 |
if len(w.shape) == 1:
|
33 |
-
w = w
|
34 |
-
|
35 |
l2 = fit_line(pts, w)
|
36 |
-
|
37 |
-
# project some points along the estimated line
|
38 |
-
p3 = l2.point_at(-10)
|
39 |
-
p4 = l2.point_at(10)
|
40 |
|
41 |
-
|
42 |
-
|
|
|
|
|
|
|
43 |
|
44 |
-
fig = plt.
|
45 |
-
|
46 |
-
|
|
|
|
|
47 |
return fig
|
48 |
|
49 |
inputs = [
|
50 |
-
gr.
|
51 |
-
gr.
|
52 |
-
gr.
|
53 |
-
gr.
|
54 |
]
|
|
|
55 |
outputs = gr.Plot()
|
56 |
|
57 |
examples = [
|
58 |
-
[0.0, 0.0,
|
59 |
]
|
60 |
|
61 |
title = 'Line Fitting'
|
@@ -69,6 +70,6 @@ demo = gr.Interface(
|
|
69 |
theme='huggingface',
|
70 |
live=True,
|
71 |
examples=examples,
|
72 |
-
|
73 |
)
|
74 |
-
|
|
|
|
|
|
1 |
import matplotlib.pyplot as plt
|
2 |
import torch
|
3 |
import matplotlib
|
4 |
matplotlib.use('Agg')
|
|
|
5 |
import gradio as gr
|
6 |
from kornia.geometry.line import ParametrizedLine, fit_line
|
7 |
|
|
|
8 |
def inference(point1, point2, point3, point4):
|
9 |
std = 1.2 # standard deviation for the points
|
10 |
num_points = 50 # total number of points
|
|
|
13 |
p0 = torch.tensor([point1, point2], dtype=torch.float32)
|
14 |
p1 = torch.tensor([point3, point4], dtype=torch.float32)
|
15 |
l1 = ParametrizedLine.through(p0, p1)
|
16 |
+
|
17 |
# sample some points and weights
|
18 |
pts, w = [], []
|
19 |
for t in torch.linspace(-10, 10, num_points):
|
|
|
22 |
p2 += p2_noise
|
23 |
pts.append(p2)
|
24 |
w.append(1 - p2_noise.mean())
|
25 |
+
|
26 |
pts = torch.stack(pts)
|
27 |
w = torch.stack(w)
|
28 |
+
|
29 |
if len(pts.shape) == 2:
|
30 |
+
pts = pts.unsqueeze(0)
|
31 |
if len(w.shape) == 1:
|
32 |
+
w = w.unsqueeze(0)
|
33 |
+
|
34 |
l2 = fit_line(pts, w)
|
|
|
|
|
|
|
|
|
35 |
|
36 |
+
# project some points along the estimated line
|
37 |
+
p3 = l2.point_at(torch.tensor(-10.0))
|
38 |
+
p4 = l2.point_at(torch.tensor(10.0))
|
39 |
+
X = torch.stack((p3, p4)).squeeze().detach().numpy()
|
40 |
+
X_pts = pts.squeeze().detach().numpy()
|
41 |
|
42 |
+
fig, ax = plt.subplots()
|
43 |
+
ax.plot(X_pts[:, 0], X_pts[:, 1], 'ro')
|
44 |
+
ax.plot(X[:, 0], X[:, 1])
|
45 |
+
ax.set_xlim(X_pts[:, 0].min() - 1, X_pts[:, 0].max() + 1)
|
46 |
+
ax.set_ylim(X_pts[:, 1].min() - 1, X_pts[:, 1].max() + 1)
|
47 |
return fig
|
48 |
|
49 |
inputs = [
|
50 |
+
gr.Slider(0.0, 10.0, value=0.0, label="Point 1 X"),
|
51 |
+
gr.Slider(0.0, 10.0, value=0.0, label="Point 1 Y"),
|
52 |
+
gr.Slider(0.0, 10.0, value=10.0, label="Point 2 X"),
|
53 |
+
gr.Slider(0.0, 10.0, value=10.0, label="Point 2 Y"),
|
54 |
]
|
55 |
+
|
56 |
outputs = gr.Plot()
|
57 |
|
58 |
examples = [
|
59 |
+
[0.0, 0.0, 10.0, 10.0],
|
60 |
]
|
61 |
|
62 |
title = 'Line Fitting'
|
|
|
70 |
theme='huggingface',
|
71 |
live=True,
|
72 |
examples=examples,
|
|
|
73 |
)
|
74 |
+
|
75 |
+
demo.launch()
|