File size: 6,588 Bytes
dadf189
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
166
167
168
169
170
171
172
173
174
175
from __future__ import annotations

from pathlib import Path
from typing import Iterable

from PIL import Image
from torchvision import transforms

from .base import FingerprintSample

IMAGE_EXTS = {".jpg", ".jpeg", ".bmp", ".png"}


class PolyULoader:
    """Loader for the PolyU Contactless-2D-to-Contact-based Fingerprint Database.

    Reference:
        Chenhao Lin, Ajay Kumar, "Matching Contactless and Contact-based
        Conventional Fingerprint Images for Biometrics Identification,"
        IEEE Transactions on Image Processing, vol. 27, pp. 2008-2021, 2018.

    Directory layout under ``root_polyu``::

        contact-based_fingerprints/
            first_session/{X}_{Y}.jpg          X=subject_id, Y=impression (1-6)
            second_session/{X}_{Y}.jpg
        processed_contactless_2d_fingerprint_images/
            first_session/p{X}/p{Y}.bmp
            second_session/p{X}/p{Y}.bmp

    The raw contactless images (1400×900, ~3.6 MB) are **not** loaded; the
    pre-downsampled grayscale versions in
    ``processed_contactless_2d_fingerprint_images`` are used instead.

    Metadata fields:
        identity_id  — ``polyu_{X}``  (shared between contact and contactless)
        finger_id    — ``f01``        (one finger per subject in this DB)
        sensor_id    — ``polyu_contact`` | ``polyu_contactless``
        dataset      — ``polyu``

    Cross-modality signal for L_sens:
        Same (identity_id, finger_id) appears under two sensor_ids, making these
        the strongest possible cross-sensor anchor pairs for GRL training.
    """

    SENSOR_CONTACT = "polyu_contact"
    SENSOR_CONTACTLESS = "polyu_contactless"
    DATASET = "polyu"
    SESSIONS = ("first_session", "second_session")

    def __init__(self, image_size: int = 224):
        self.transform = transforms.Compose(
            [
                transforms.Resize((image_size, image_size)),
                transforms.ToTensor(),
            ]
        )

    def discover(self, root_polyu: str) -> list[dict[str, str]]:
        """Scan the PolyU root directory and return a list of metadata records.

        Args:
            root_polyu: Path to the PolyU root folder (the folder containing
                ``contact-based_fingerprints`` and
                ``processed_contactless_2d_fingerprint_images``).

        Returns:
            List of dicts with keys: image_path, identity_id, finger_id,
            sensor_id, dataset.
        """
        root = Path(root_polyu)
        if not root.exists():
            raise FileNotFoundError(f"PolyU root not found: {root}")

        records: list[dict[str, str]] = []
        records.extend(self._scan_contact(root))
        records.extend(self._scan_contactless(root))
        return records

    def iter_samples(
        self, records: Iterable[dict[str, str]]
    ) -> Iterable[FingerprintSample]:
        for rec in records:
            image = Image.open(rec["image_path"]).convert("L")
            tensor = self.transform(image)
            yield {
                "image": tensor,
                "identity_id": rec["identity_id"],
                "finger_id": rec["finger_id"],
                "sensor_id": rec["sensor_id"],
                "dataset": rec["dataset"],
                "image_path": rec["image_path"],
            }

    # ------------------------------------------------------------------
    # Private helpers
    # ------------------------------------------------------------------

    def _scan_contact(self, root: Path) -> list[dict[str, str]]:
        """Scan contact-based_fingerprints/{session}/{X}_{Y}.jpg."""
        folder = root / "contact-based_fingerprints"
        records: list[dict[str, str]] = []
        if not folder.exists():
            return records
        for session in self.SESSIONS:
            session_dir = folder / session
            if not session_dir.exists():
                continue
            for img_path in sorted(session_dir.iterdir()):
                if not img_path.is_file():
                    continue
                if img_path.suffix.lower() not in IMAGE_EXTS:
                    continue
                meta = self._parse_contact_filename(img_path)
                if meta is not None:
                    records.append(meta)
        return records

    def _scan_contactless(self, root: Path) -> list[dict[str, str]]:
        """Scan processed_contactless_2d_fingerprint_images/{session}/p{X}/p{Y}.bmp."""
        folder = root / "processed_contactless_2d_fingerprint_images"
        records: list[dict[str, str]] = []
        if not folder.exists():
            return records
        for session in self.SESSIONS:
            session_dir = folder / session
            if not session_dir.exists():
                continue
            # Each subject has a subdirectory named p{X}
            for subj_dir in sorted(session_dir.iterdir()):
                if not subj_dir.is_dir():
                    continue
                subj_name = subj_dir.name  # e.g. "p42"
                if not subj_name.lower().startswith("p"):
                    continue
                try:
                    subject_id = str(int(subj_name[1:]))
                except ValueError:
                    continue
                identity_id = f"polyu_{subject_id}"
                for img_path in sorted(subj_dir.iterdir()):
                    if not img_path.is_file():
                        continue
                    if img_path.suffix.lower() not in IMAGE_EXTS:
                        continue
                    records.append(
                        {
                            "image_path": str(img_path),
                            "identity_id": identity_id,
                            "finger_id": "f01",
                            "sensor_id": self.SENSOR_CONTACTLESS,
                            "dataset": self.DATASET,
                        }
                    )
        return records

    @staticmethod
    def _parse_contact_filename(img_path: Path) -> dict[str, str] | None:
        """Parse '{X}_{Y}.jpg' → identity_id='polyu_{X}'."""
        stem = img_path.stem  # e.g. "42_3"
        parts = stem.split("_")
        if len(parts) != 2:
            return None
        try:
            subject_id = str(int(parts[0]))
        except ValueError:
            return None
        return {
            "image_path": str(img_path),
            "identity_id": f"polyu_{subject_id}",
            "finger_id": "f01",
            "sensor_id": PolyULoader.SENSOR_CONTACT,
            "dataset": PolyULoader.DATASET,
        }