PuzzleProject2021/puzzle.py at main · firluk/PuzzleProject2021 · 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
162
163
164
165
import matplotlib.pyplot as plt
import numpy as np
import skimage

from facet import Facet, compatibility_func
from piece import Piece, pieces_from_masks
from puzzle_piece_detector.inference_callable import Inference
from utils import image_in_scale, masks_in_scale


def evaluate_edge_compatibility(pieces, comparison_method):
    n_side_pieces = n_middle_pieces = n_corner_pieces = 0
    for piece in pieces:
        if piece.type is Piece.Type.SIDE:
            n_side_pieces = n_side_pieces + 1
        elif piece.type is Piece.Type.MIDDLE:
            n_middle_pieces = n_middle_pieces + 1
        elif piece.type is Piece.Type.CORNER:
            n_corner_pieces = n_corner_pieces + 1
        else:
            raise
    n_pieces = len(pieces)
    n_facets = 4
    arr = np.zeros((n_pieces, n_pieces, n_facets, n_facets))
    for p1idx, p1 in enumerate(pieces):
        for p2idx, p2 in enumerate(pieces):
            if p1idx < p2idx:
                for f1idx, f1 in enumerate(p1.facets):
                    if f1.type is Facet.Type.FLAT:
                        continue
                    for f2idx, f2 in enumerate(p2.facets):
                        if f2.type is Facet.Type.FLAT:
                            continue
                        arr[p1idx, p2idx, f1idx, f2idx] = comparison_method(f1, f2)
    return arr, n_facets, n_pieces, n_side_pieces, n_middle_pieces


def segment_to_masks_and_extract_pieces(weights_path, image_path, segmenting_method):
    if segmenting_method[0]:
        inference = Inference(weights_path)
        masks = segmenting_method[1](inference, image_path)
    else:
        masks = segmenting_method[1](image_path)
    image = skimage.io.imread(image_path)
    scale = 1
    masks = masks_in_scale(masks, scale)
    image = image_in_scale(image, scale)
    pieces = pieces_from_masks(masks, image)
    return pieces, masks


def print_pieces(pieces, name):
    fig = plt.figure()
    for i in range(len(pieces)):
        factor = int(np.ceil(np.sqrt(len(pieces))))
        ax = fig.add_subplot(factor, factor, i + 1)
        plt.imshow(pieces[i].cropped_image)
        ax.set_title(f'{i}')
        ax.axis('off')
    plt.savefig(f'plots/{name}.png')
    plt.close(fig)


def print_facets(pieces, name):
    fig = plt.figure()
    for i in range(len(pieces)):
        factor = int(np.ceil(np.sqrt(len(pieces))))
        ax = fig.add_subplot(factor, factor, i + 1)
        img = np.zeros_like(pieces[i].cropped_image)
        facet_colors = np.array([[255, 0, 0], [0, 255, 0], [0, 0, 255], [255, 255, 255]]).astype(np.uint8)
        for fi in range(len(pieces[i].facets)):
            mask = pieces[i].facets[fi].facet_mask
            img[mask, :] = facet_colors[fi, :]
        plt.imshow(img)
        ax.set_title(f'{i}')
        ax.axis('off')

    plt.savefig(f'plots/{name}.png')
    plt.close(fig)


def print_figures_with_weights_to_folder(edges, arr, pieces, output_dir):
    for edge in edges:
        p1, p2, f1, f2 = edge[0], edge[1], edge[2], edge[3]

        fig = plt.figure()

        ax = fig.add_subplot(2, 2, 1)
        plt.imshow(pieces[p1].cropped_image)
        ax.set_title(f'{p1}')
        ax = fig.add_subplot(2, 2, 2)
        plt.imshow(pieces[p2].cropped_image)
        ax.set_title(f'{p2}')
        ax = fig.add_subplot(2, 2, 3)
        plt.imshow(pieces[p1].facets[f1].facet_mask)
        ax.set_title(f'{f1}')
        ax = fig.add_subplot(2, 2, 4)
        plt.imshow(pieces[p2].facets[f2].facet_mask)
        ax.set_title(f'{f2}')

        plt.savefig(f'plots/{output_dir}/{format(arr[p1, p2, f1, f2], ".3f")}_{p1}_{p2}_{f1}_{f2}.png')
        plt.close()


def sort_and_filter(n_pieces, n_facets, filter_val, weights, descending=True):
    sort_idx = np.argsort(weights, axis=None)
    if descending:
        sort_idx = np.flip(sort_idx)  # descending order using flat
    mask = weights.flat > filter_val  # filter mask,
    filtered_idx = sort_idx[mask[sort_idx]]  # cutting according to filter
    # get indices
    edges = np.transpose(np.vstack(np.unravel_index(filtered_idx, (n_pieces, n_pieces, n_facets, n_facets))))
    return edges


def paint_facets_by_type(masks, pieces):
    width, height, _ = masks.shape
    masks_with_facets = np.ones((width, height, 3), dtype=np.uint8) * 255
    for piece in pieces:
        img = np.ones_like(piece.cropped_image) * 255
        facet_colors = dict()
        facet_colors[Facet.Type.FLAT] = np.array([0, 0, 0], dtype=np.uint8)
        facet_colors[Facet.Type.TAB] = np.array([0, 255, 0], dtype=np.uint8)
        facet_colors[Facet.Type.BLANK] = np.array([255, 0, 0], dtype=np.uint8)
        for fi in range(len(piece.facets)):
            facet = piece.facets[fi]
            mask = facet.facet_mask_custom_contour_size(3)
            img[mask, :] = facet_colors[facet.type]
        height, width, _ = piece.cropped_image.shape
        left, top = piece.left, piece.top
        left_width, top_height = left + width, top + height
        masks_with_facets[top:top_height, left:left_width] = img
    return masks_with_facets


def paint_facets_distinct(masks, pieces):
    width, height, _ = masks.shape
    masks_with_facets = np.ones((width, height, 3), dtype=np.uint8) * 255
    for piece in pieces:
        img = np.ones_like(piece.cropped_image) * 255
        facet_colors = np.array([[255, 0, 0], [0, 255, 0], [0, 0, 255], [0, 0, 0]]).astype(np.uint8)
        for fi in range(len(piece.facets)):
            mask = piece.facets[fi].facet_mask_custom_contour_size(3)
            img[mask, :] = facet_colors[fi, :]
        height, width, _ = piece.cropped_image.shape
        left, top = piece.left, piece.top
        left_width, top_height = left + width, top + height
        masks_with_facets[top:top_height, left:left_width] = img
    return masks_with_facets


def calc_cmp_from_iou_and_mgc(iou, mgc, n_facets, n_pieces, pieces):
    cmp = np.zeros((n_pieces, n_pieces, n_facets, n_facets))
    for p1idx, p1 in enumerate(pieces):
        for p2idx, p2 in enumerate(pieces):
            if p1idx < p2idx:
                for f1idx, f1 in enumerate(p1.facets):
                    if f1.type is Facet.Type.FLAT:
                        continue
                    for f2idx, f2 in enumerate(p2.facets):
                        if f2.type is Facet.Type.FLAT:
                            continue
                        cmp[p1idx, p2idx, f1idx, f2idx] = compatibility_func()(mgc[p1idx, p2idx, f1idx, f2idx],
                                                                               iou[p1idx, p2idx, f1idx, f2idx])
    return cmp