ImmobileParticles_MatchAndAnalyse/testImageAlign.py at main · robna/ImmobileParticles_MatchAndAnalyse · 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
155
156
157
158
159
160
161
import unittest
import numpy as np
import alignImages as ai


class TestImageAlign(unittest.TestCase):
    def assert_array_almost_equal(self, arr1: np.ndarray, arr2: np.ndarray, places: int = 7) -> None:
        self.assertEqual(arr1.shape, arr2.shape)
        arr1, arr2 = arr1.ravel(), arr2.ravel()
        for elem1, elem2 in zip(arr1, arr2):
            self.assertAlmostEqual(elem1, elem2, places=places)

    def test_get_error_from_centers(self):
        points: np.ndarray = np.arange(10).reshape((5, 2))
        errors, indBefore2After = ai.getIndicesAndErrosFromCenters(points, points)
        self.assertEqual(np.sum(errors), 0.0)
        for key, val in indBefore2After.items():
            self.assertEqual(key, val)

        # offset points by a bit
        points2: np.ndarray = points + np.array([0.01, 0])
        errors, indBefore2After = ai.getIndicesAndErrosFromCenters(points, points2)

        for key, val in indBefore2After.items():
            self.assertEqual(key, val)
        self.assertAlmostEqual(np.sum(errors), 0.05)

        pointsCopy = points.copy()
        # move out first point too far
        pointsCopy[0, :] += np.array([5, 0])
        # also, add a list of max Errors per particle that are being used as distance tolerance for the matching:
        maxErrors = [5, 0, 0, 0, 0]  # first point is allowed to be max 5 units away
        errors, indBefore2After = ai.getIndicesAndErrosFromCenters(points, pointsCopy, maxErrors)
        self.assertEqual(np.sum(errors), 5)  # Result is fine

        maxErrors = [0, 5, 0, 0, 0]  # second point is allowed to be max 5 units away, but not the first
        errors, indBefore2After = ai.getIndicesAndErrosFromCenters(points, pointsCopy, maxErrors)
        self.assertEqual(np.sum(errors), 100000)  # the 100000 is added as high penalty, as the first point is too far away
        self.assertEqual(len(indBefore2After), 4)  # and only four particles are matched
        self.assertTrue(0 not in indBefore2After.keys())  # and particle 0 is not in the dictionary

        maxErrors = [4.9, 0, 0, 0, 0]  # Tolerance for first point just below actual distance
        errors, indBefore2After = ai.getIndicesAndErrosFromCenters(points, pointsCopy, maxErrors)
        self.assertEqual(np.sum(errors), 100000)  # the 100000 is added as high penalty, as the first point is too far away
        self.assertEqual(len(indBefore2After), 4)  # and only four particles are matched
        self.assertTrue(0 not in indBefore2After.keys())  # and particle 0 is not in the dictionary

        points2Copy = points2.copy()
        # let's swap some of the points
        points2[0, :], points2[1, :] = points2Copy[1, :], points2Copy[0, :]
        points2[2, :], points2[4, :] = points2Copy[4, :], points2Copy[2, :]
        errors, indBefore2After = ai.getIndicesAndErrosFromCenters(points, points2)
        expectedInd = {0: 1,
                       1: 0,
                       2: 4,
                       3: 3,
                       4: 2}

        self.assertAlmostEqual(np.sum(errors), 0.05)
        self.assertEqual(len(indBefore2After), 5)
        for key in indBefore2After.keys():
            self.assertEqual(indBefore2After[key], expectedInd[key])

        # lets add distance tolerance to that...
        maxErrors = [0, 0.1, 0.1, 0.1, 0.1]  # first point should be thrown out as it only allows for 0 error.
        errors, indBefore2After = ai.getIndicesAndErrosFromCenters(points, points2, maxErrors)
        self.assertAlmostEqual(np.sum(errors), 100000.04)
        self.assertTrue(0 not in indBefore2After.keys())


        # add points that are outliers
        points3: np.ndarray = np.zeros((7, 2))
        points3[:5, :] = points2
        points3[5:, :] = np.array([[10, 20], [-2.5, 3.5]])
        errors, indBefore2After = ai.getIndicesAndErrosFromCenters(points, points3)
        self.assertAlmostEqual(np.sum(errors), 0.05)
        for key in indBefore2After.keys():
            self.assertEqual(indBefore2After[key], expectedInd[key])

        points4: np.ndarray = points2[:3, :]  # let's have less points in the second array
        errors, indBefore2After = ai.getIndicesAndErrosFromCenters(points, points4)
        self.assertAlmostEqual(np.sum(errors), 0.03)
        expectedInd = {1: 0,
                       0: 1,
                       4: 2}
        for key in indBefore2After.keys():
            self.assertEqual(indBefore2After[key], expectedInd[key])

    def test_offset_points(self):
        points: np.ndarray = np.arange(10).reshape((5, 2))
        # only offsets
        for shiftX in [-5, 0, 5]:
            for shiftY in [-3, 0, 3]:
                shift: np.ndarray = np.array([shiftX, shiftY])
                newPoints: np.ndarray = ai.offSetPoints(points, 0, shift)
                for i in range(points.shape[0]):
                    self.assertEqual(points[i, 0] + shift[0], newPoints[i, 0])
                    self.assertEqual(points[i, 1] + shift[1], newPoints[i, 1])

        # only angles:
        shift = np.array([0, 0])
        points = np.array([[0, 1], [0, -1]])
        angle = 90
        newPoints = ai.offSetPoints(points, angle, shift)
        self.assert_array_almost_equal(newPoints, np.array([[-1, 0], [1, 0]]))

        angle = 180
        newPoints = ai.offSetPoints(points, angle, shift)
        self.assert_array_almost_equal(newPoints, np.array([[0, -1], [0, 1]]))

        angle = 90
        points = np.array([[1, 1], [-1, -1]])
        newPoints = ai.offSetPoints(points, angle, shift)
        self.assert_array_almost_equal(newPoints, np.array([[-1, 1],  [1, -1]]))

        # rot and offset:
        shift = [2, 1]
        newPoints = ai.offSetPoints(points, angle, shift)
        self.assert_array_almost_equal(newPoints, np.array([[1, 2], [3, 0]]))

    def test_get_offset_and_rot(self):
        points: np.ndarray = np.array([[1, 1], [1, 0], [0, 2], [1, 2]])
        shifts: list = [[0, 0], [1, 0], [1, 1], [-1, -1]]
        angles: list = [-20, -10, 0, 10, 20]
        for shift in shifts:
            for angle in angles:
                newPoints: np.ndarray = ai.offSetPoints(points, angle, np.array(shift))
                optAngle, optShift = ai.findAngleAndShift(points, newPoints)
                self.assertAlmostEqual(optAngle, angle, places=3)
                self.assertAlmostEqual(optShift[0], shift[0], places=5)
                self.assertAlmostEqual(optShift[1], shift[1], places=5)

                transformedPoints: np.ndarray = ai.offSetPoints(points, optAngle, optShift)
                err, _ = ai.getIndicesAndErrosFromCenters(newPoints, transformedPoints)
                self.assertAlmostEqual(np.sum(err), 0, places=3)
                self.assert_array_almost_equal(transformedPoints, newPoints)

                newPoints2 = newPoints.copy()
                newPoints2[2, :], newPoints2[0, :] = newPoints[0, :], newPoints[2, :]  # switch indices
                optAngle, optShift = ai.findAngleAndShift(points, newPoints2)
                transformedPoints = ai.offSetPoints(points, optAngle, optShift)
                err, _ = ai.getIndicesAndErrosFromCenters(transformedPoints, newPoints2)
                self.assertAlmostEqual(np.sum(err), 0, places=6)
                self.assertAlmostEqual(optAngle, angle, places=6)
                self.assertAlmostEqual(optShift[0], shift[0])
                self.assertAlmostEqual(optShift[1], shift[1])

                newPoints3 = np.zeros((newPoints.shape[0] + 2, 2))
                newPoints3[:newPoints2.shape[0], :] = newPoints2
                newPoints3[newPoints2.shape[0]:, :] = np.array([[-100, 100], [100, 100]])
                optAngle, optShift = ai.findAngleAndShift(points, newPoints3)
                transformedPoints: np.ndarray = ai.offSetPoints(points, optAngle, optShift)
                err, _ = ai.getIndicesAndErrosFromCenters(transformedPoints, newPoints3)
                self.assertAlmostEqual(np.sum(err), 0, places=6)
                self.assertAlmostEqual(optAngle, angle, places=6)
                self.assertAlmostEqual(optShift[0], shift[0])
                self.assertAlmostEqual(optShift[1], shift[1])


if __name__ == '__main__':
    unittest.main()