Generalized Procrustes analysis from TPS file

import numpy as np
import pandas as pd

import matplotlib.pyplot as plt
import seaborn as sns

from sklearn.decomposition import PCA

from ktch.landmark import GeneralizedProcrustesAnalysis
from ktch.io import read_tps

Reading TPS file

df_triangles = read_tps("./landmarks_triangle.tps", as_frame=True)
df_triangles

		x	y
specimen_id	coord_id
0	0	-0.531046	-0.280265
	1	-0.177303	1.107969
	2	0.850131	1.064210
1	0	-0.480121	-0.145592
	1	-0.172134	0.871833
...	...	...	...
48	1	-0.153270	0.980145
	2	0.692266	0.797412
49	0	0.286409	-0.353412
	1	0.213187	0.819529
	2	0.905297	1.175739

150 rows × 2 columns

fig, ax = plt.subplots()
sns.scatterplot(
    data = df_triangles,
    x="x",y="y", 
    hue="specimen_id", style="coord_id",ax=ax
    )
ax.set_aspect('equal')
ax.legend(loc="upper left", bbox_to_anchor=(1, 1))

<matplotlib.legend.Legend at 0x7fc358d36f50>

GPA

gpa = GeneralizedProcrustesAnalysis().set_output(transform="pandas")

df_coords = df_triangles.unstack().swaplevel(1, 0, axis=1).sort_index(axis=1)
df_coords.columns = [dim +"_"+ str(landmark_idx) for landmark_idx,dim in df_coords.columns]
df_coords

df_shapes = gpa.fit_transform(df_coords)
df_shapes

	x_0	y_0	x_1	y_1	x_2	y_2
specimen_id
0	-0.308613	-0.638885	-0.182910	0.294746	0.491523	0.344139
1	-0.372183	-0.487429	-0.196928	0.072805	0.569111	0.414624
10	-0.397825	-0.561788	-0.035101	0.124282	0.432926	0.437506
11	-0.373016	-0.613275	-0.047044	0.202708	0.420060	0.410568
12	-0.252899	-0.620448	-0.273916	0.334577	0.526814	0.285871
13	-0.254792	-0.594593	-0.302705	0.305684	0.557497	0.288908
14	-0.382952	-0.569840	0.082589	0.152464	0.300364	0.417376
15	-0.219393	-0.552444	-0.364955	0.300376	0.584348	0.252068
16	-0.320267	-0.494096	-0.293551	0.133416	0.613818	0.360679
17	-0.339855	-0.603090	-0.181716	0.224619	0.521571	0.378471
18	-0.164839	-0.213845	-0.460237	0.018226	0.625077	0.195619
19	-0.210433	-0.539583	-0.379697	0.296870	0.590129	0.242713
2	-0.366486	-0.605645	-0.106602	0.200330	0.473088	0.405315
20	-0.258150	-0.504268	-0.364335	0.209576	0.622485	0.294692
21	-0.277794	-0.566510	-0.306544	0.252170	0.584337	0.314340
22	-0.225401	-0.676491	-0.160386	0.424549	0.385786	0.251942
23	-0.258153	-0.676369	-0.146719	0.388880	0.404872	0.287489
24	-0.280671	-0.525416	-0.330592	0.207091	0.611263	0.318325
25	-0.215296	-0.505604	-0.397853	0.256922	0.613148	0.248682
26	-0.298502	-0.529983	-0.307048	0.192828	0.605550	0.337154
27	-0.278224	-0.639372	-0.222512	0.327351	0.500736	0.312021
28	-0.279808	-0.576080	-0.296562	0.259860	0.576370	0.316220
29	-0.338928	-0.453639	-0.273890	0.073037	0.612818	0.380602
3	-0.322719	-0.611769	-0.202282	0.251475	0.525002	0.360294
30	-0.274653	-0.499217	-0.349916	0.186850	0.624569	0.312367
31	-0.298402	-0.477670	-0.329793	0.139829	0.628195	0.337841
32	-0.176246	-0.605878	-0.325364	0.402554	0.501610	0.203323
33	-0.172484	-0.478247	-0.435118	0.275371	0.607602	0.202876
34	-0.210254	-0.662374	-0.213811	0.425340	0.424065	0.237034
35	-0.361575	-0.629273	-0.000713	0.232784	0.362288	0.396489
36	-0.326016	-0.656912	-0.093636	0.296542	0.419652	0.360370
37	-0.219476	-0.656608	-0.228225	0.408971	0.447701	0.247637
38	-0.215303	-0.490933	-0.405267	0.241987	0.620570	0.248947
39	-0.191623	-0.543638	-0.385946	0.321395	0.577570	0.222243
4	-0.305088	-0.667007	-0.110755	0.329094	0.415843	0.337913
40	-0.180727	-0.544012	-0.389820	0.333616	0.570547	0.210395
41	-0.292490	-0.527475	-0.315843	0.196640	0.608334	0.330835
42	-0.339802	-0.505129	-0.257392	0.124161	0.597194	0.380967
43	-0.280448	-0.667827	-0.154245	0.355591	0.434693	0.312236
44	-0.152263	-0.556959	-0.384438	0.378026	0.536701	0.178933
45	-0.188162	-0.585891	-0.348026	0.368723	0.536188	0.217168
46	-0.097880	-0.420235	-0.481143	0.297774	0.579024	0.122461
47	-0.143493	-0.576295	-0.363271	0.407702	0.506765	0.168593
48	-0.085252	-0.549410	-0.377222	0.444325	0.462474	0.105085
49	-0.337458	-0.635395	-0.130722	0.261236	0.468180	0.374159
5	-0.272885	-0.612977	-0.265891	0.305395	0.538776	0.307582
6	-0.318673	-0.540378	-0.273450	0.182154	0.592124	0.358224
7	-0.302482	-0.635783	-0.197969	0.297893	0.500451	0.337890
8	-0.231492	-0.657091	-0.223173	0.396416	0.454665	0.260675
9	-0.162869	-0.625502	-0.292318	0.437962	0.455187	0.187540

df_shapes_vis = df_shapes.copy()
df_shapes_vis.columns = pd.MultiIndex.from_tuples([[int(landmark_idx), dim] for dim, landmark_idx in [idx.split("_") for idx in df_shapes_vis.columns]], names=["coord_id","dim"])
df_shapes_vis.sort_index(axis=1, inplace=True)
df_shapes_vis = df_shapes_vis.swaplevel(0,1,axis=1).stack(level=1)
df_shapes_vis

/tmp/ipykernel_2204/3231015902.py:4: FutureWarning: The previous implementation of stack is deprecated and will be removed in a future version of pandas. See the What's New notes for pandas 2.1.0 for details. Specify future_stack=True to adopt the new implementation and silence this warning.
  df_shapes_vis = df_shapes_vis.swaplevel(0,1,axis=1).stack(level=1)

	dim	x	y
specimen_id	coord_id
0	0	-0.308613	-0.638885
	1	-0.182910	0.294746
	2	0.491523	0.344139
1	0	-0.372183	-0.487429
	1	-0.196928	0.072805
...	...	...	...
8	1	-0.223173	0.396416
	2	0.454665	0.260675
9	0	-0.162869	-0.625502
	1	-0.292318	0.437962
	2	0.455187	0.187540

150 rows × 2 columns

fig, ax = plt.subplots()
sns.scatterplot(
    data = df_shapes_vis,
    x="x",y="y", 
    hue="specimen_id", style="coord_id",ax=ax
    )
ax.set_aspect('equal')
ax.legend(loc="upper left", bbox_to_anchor=(1, 1))

<matplotlib.legend.Legend at 0x7fc3568c6f50>