mirror of
https://github.com/wiltodelta/remove-ai-watermarks.git
synced 2026-10-14 13:30:55 +02:00
* Add rigorous SynthID research and evaluation harness * Record SynthID cross-epoch validation results * Reset SynthID detector hypotheses to paired training * Add paired wavelet and spectral SynthID probes * Record closed-corpus SynthID detector ablations * Add registered SynthID phase probing * Add periodic SynthID tile probing * Add calibrated SynthID pixel detector * Harden scale-registered SynthID detection * Record SynthID phase and channel rescue findings * Add production SynthID routing and OpenAI verification * Reframe periodic pixel route as pipeline lattice; confirm and harden detection The frozen periodic experts read an origin-anchored generation-pipeline lattice destroyed by a crop off the tile grid, not the crop-robust SynthID mark. Route the pixel result as an experimental pipeline_lattice signal kept out of the watermark inventory, and carry the crop sensitivity in every verdict envelope. Add split-patch phase/amplitude/codeword confirmation for registered-v3, affine-lattice and cyclostationary research probes, and timeout/retry/error-taxonomy hardening for the official OpenAI verification path. * Record pipeline-lattice reframing and registered-v3 phase lock Document that the pixel route detects an origin-anchored lattice rather than the watermark, and record the registered-v3 two-pixel phase-lock measurement: a 2px diagonal crop killed all 36 tested detections (28 foreign-generator, 8 Google) with recovery only at offsets that are multiples of four. Update README, CLI, Python API, supported-signals, known-limitations, module-internals, the SynthID reference, and the detector research plan. * Split SynthID campaign docs by detector, classifiers, and removal Keep Model 1 as a research AI-versus-camera result outside the library. Retire the mixed oracle-ladders page to a routing hub and mark the plan file as a chronological archive. * Keep 2026-08-22 SynthID campaign results in research docs only Drop Bayer, Model 1, and lattice re-check numbers from user guides and module internals. The library surface is unchanged. * Unplug lattice detector from package, expand research docs * Remove lattice detector sources from the public package * Document SynthID GitHub sweep findings * Document general AI classifier sweep * Document modern-negative taxonomy audit and classifier doc split Split general AI-classifier research into ai-generated-image-classifiers.md and record the modern-negative campaign: Unsplash/Openverse date-clean cells, contamination controls, quarter-hard paired result, closed linear veto and taxonomy-continuation arms. pre-commit: 1) maintain.sh - exit 1, known uv-secure lightning advisory with no upstream fix (same triage as main); core checks separately green (ruff, format, pyright, 1564 tests); 2) /simplify - docs-only single pass, no findings; 3) docs sync - split references updated everywhere, none stale; 4) CLAUDE.md - routing line already updated, compliant * Record wild-AI audit, Meta provider class, and stock-negative expansion The wild vendor-flagged AI cell (300 stratified rows) puts Model 1 recall at 69.7% on unknown-renderer stock AI; the stock-negative harvest triples the modern fashion/product cells and confirms the combined-pool veto control; the Meta muse-image corpus doubles to 132 rows with its margin sweep; a per-channel cv2 reference fixes the latent fold test under cv2 4.10.0. pre-commit: 1) maintain.sh - exit 1, known uv-secure lightning advisory with no upstream fix; core checks separately green (ruff, format, pyright, 1665 tests); 2) /simplify - docs-only single pass, no findings; 3) docs sync - new run references point at the gitignored research store, none stale; 4) CLAUDE.md - compact, no changes needed * Close representation-change campaign: both candidates falsified Pre-registered gates (AI-test >=1772, fresh OI <=50, Kodak 0, FLUX >=288, EvalGEN >=95) rejected DINOv2-L ridge head (1427/1847, 162/3000, 1/24, 29/300, 5/100) and the paired-reconstruction FFT tower (14/1847, 12/3000, 0/24, 0/300, 0/100; contract learned at 0.949 pairwise separation but no transfer to production generators). Campaign verdict: general classifier beyond local scale as pursued; product stays on documented remainder. pre-commit: 1) maintain.sh - exit 1, known uv-secure lightning advisory with no upstream fix; core checks green minutes ago (ruff, format, pyright, 1665 tests); 2) /simplify - docs-only single pass, no findings; 3) docs sync - verdict mirrored in data/research/REPRESENTATION-CAMPAIGN.md, no stale refs; 4) CLAUDE.md - no changes * Record full failure map, H2 router falsification, and H3 process detector closure 20 measured negative domains with batch-2 (memes, whiteboards, tickets, book pages). H2 domain router falsified: 97.9% of AI-test looks structured, every family 94-100%. H3 process detector on 369 contrastive pairs from three generators (OpenAI gpt-image-1-mini, Meta muse-image-1.0, Gemini flash-image) achieves 0.909 pairwise AUC on training pairs but transfers at AUC 0.55-0.60 to production AI; human receipts score higher than any AI family. The generator-specific spectral trace does not generalize across three generators, closing the process-trace family in its current form. pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (ruff, format, pyright, 1665 tests); 2) /simplify - docs + research scripts, single pass, no findings; 3) docs sync - all artifacts in data/research/, no stale refs; 4) CLAUDE.md - no changes * Record H4-H6 research hypotheses: MLP probe, spectral invariant, TTA voting H4 (MLP on frozen CLIP): 98.6% AI-test recall but 15.8% photo FPR — same trade-off curve as quarter-hard. H5 (spectral invariant from CVPR 2025): 2.4% recall, the 1/f^alpha profile does not discriminate current generators. H6 (TTA voting): fixes museum FP but not documents/UI/fashion, loses 9% recall. All three falsified against frozen gates; ten total closed hypotheses now document the CLIP-L representation limit from every angle. Also records the provider cascade across all 20 domains (69.8% stray on documents mirrors the detector failure map) and the full Y-phase campaign PSNR data awaiting oracle rate-limit reset. pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (ruff, format, pyright, 1665 tests); 2) /simplify - docs-only single pass, no findings; 3) docs sync - all artifacts in data/research/, no stale refs; 4) CLAUDE.md - no changes * Fix Y-phase frequency convention, record H4-H6 and campaign results The initial Y-phase campaign targeted FFT bins 16-32 instead of spatial periods 16-32 px (1/32 to 1/16 cycles/px). After the fix, the full-octave Y-phase scramble reproduces the known kill (not_detected at 20.5 dB on s1 at native 1122x1402). Partial sub-bands are visually invisible (42.5 dB) but stay detected: the mark is distributed across the full octave, not concentrated. Point intervention is not viable within this frequency band. H4 MLP probe: 98.6% recall, 15.8% FPR (same trade-off wall). H5 spectral invariant: 2.4% recall (1/f^alpha does not discriminate). H6 TTA voting: fixes museum FP but not documents/UI/fashion. pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (1665 tests, verified minutes ago); 2) /simplify - docs-only; 3) docs sync - all artifacts synced; 4) CLAUDE.md - no changes * Add TC260 fifth class and per-class margin calibration to provider cascade Five-head cascade (openai/google/tc260/meta/no_ai): openai 68.8%, google 76.6%, tc260 63.8%, meta 71.8% mean recall at uniform margin 0.50. Per-class margins calibrated against 1.67% photo budget: openai 0.31, google 0.29, tc260 0.39, meta 0.55. TC260 test recall 71.4% at calibrated margins on 196 held-out images. Chinese-AIGC is a viable pixel class. pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (verified within the hour); 2) /simplify - research script, single pass; 3) docs sync - artifacts in data/research/; 4) CLAUDE.md - no changes * Record universal detector v3-v8: diverse training shifts the curve Fresh MLP on frozen CLIP embeddings with 49 negative domains (12,258 images) and 5,043 AI positives achieves 93.0% AI-test recall with 39/49 domains at zero false positives. Digital art FP dropped from 99.3% (Model 1) to 24.8%, fashion from 62-71% to 12%, UI from 99.3% to 12%. Photo FPR 0.7%. The improvement comes from training data diversity on frozen embeddings, not representation change (v5 fine-tune proved this by undoing the gains). Also adds .local-eval to ruff exclude so the gate covers only tracked files. pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (ruff, format, pyright, 1731 tests); 2) /simplify - docs + research; 3) docs sync - all artifacts in data/research/; 4) CLAUDE.md - no changes * Add three-tier ensemble detector: DEFINITELY/POSSIBLY/LIKELY_HUMAN Combines Model 1 and v8 into calibrated confidence tiers. DEFINITELY_AI (both models agree) achieves 79.6% recall at 0.50% photo FPR with 46/49 domains at zero FP. The union reaches 93.3% recall at 1.80% photo FPR with Kodak 0/24. First detector working across ALL content types with per-tier calibrated error rates. pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (ruff, format, pyright, 1731 tests); 2) /simplify - docs + ensemble; 3) docs sync - artifacts in data/research/; 4) CLAUDE.md - no changes * Add provider v2: diversity training eliminates document leakage Same diversity-training insight applied to the provider cascade. Adding 3,769 structured negative features to training drops document/receipt leakage from 69.8% to 5.1%, UI from 10% to 4.3%, fashion from 5-15% to 0-3%. AI recall improves: OpenAI 84% (was 75%), Google 88.8% (was 79%). Photo leakage 1.4%. Worst regression: memes 47.9% (was 27%). pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (ruff, format, pyright, 1731 tests); 2) /simplify - provider v2; 3) docs sync - artifacts in data/research/; 4) CLAUDE.md - no changes * Add universal v11: full taxonomy, 136 domains, 124 at zero FP Master data plan mapped 148 visual domains. Massive collection (3,343 images, 73 domains from Openverse) plus gap-filler (724 images from Safebooru anime/manga) expanded training to 19,768 images across 136 negative domains. Universal v11 (MLP-512 + Model 1 ensemble): DEFINITELY 90.3% recall at 0.50% photo FPR, ANY 95.6% recall, 124/134 domains at exactly 0% FP, worst FP 13.8% (digital art). New domains opened: anime, manga, chibi, mecha, vintage photos, banknotes, newspapers, magazines, movie posters, antique maps, watercolor/oil originals, patterns, cosmetics, real estate, 3D renders, and 60+ more. pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (ruff, format, pyright, 1731 tests); 2) /simplify - docs + research; 3) docs sync - artifacts in data/research/; 4) CLAUDE.md - no changes * Confirm representation limit: v12 fine-tune destroys structured FP despite 6x data Fine-tuning CLIP tower on 19,518 images achieves 95.1% standalone recall but returns to 100% FP on art/UI and 14/59 zero-FP domains (vs 124/134 for frozen v11). This is a representation property, not a data issue. Frozen CLIP + MLP (v11) is the only working architecture. Also records provider v3 trained on all negative domains. pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (ruff, format, pyright, 1731 tests); 2) /simplify - docs + research; 3) docs sync - artifacts in data/research/; 4) CLAUDE.md - no changes * Document v13 plateau: 128/136 zero FP, frozen-CLIP ceiling confirmed v13 with top-plan expansion (1,169 new images) improves domain coverage to 128/136 at zero FP and reduces fashion FP to 5.3%, but DEFINITELY recall plateaus at 89.6% and art FP at 15.9%. The frozen CLIP + MLP architecture has reached its ceiling. pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (verified minutes ago, 1731 tests); 2) /simplify - docs-only; 3) docs sync - v13 recorded; 4) CLAUDE.md - no changes * Document the 2026-08-31 photo-classifier freeze. Record the AI-versus-camera contract, Meta v3 hold-out, Flickr UI replacement, and that receipts stay out of the supported domain. * Scrub Spaces catalog details from classifier research docs * Record receipt layout-gate probe for the universal detector
183 lines
7.0 KiB
Python
183 lines
7.0 KiB
Python
"""Synthetic tests for the paired SynthID spectral research harness."""
|
|
|
|
from __future__ import annotations
|
|
|
|
import sys
|
|
from itertools import combinations
|
|
from pathlib import Path
|
|
|
|
import numpy as np
|
|
from PIL import Image
|
|
|
|
sys.path.insert(0, str(Path(__file__).resolve().parent.parent / "scripts"))
|
|
|
|
import synthid_spectral_probe as probe
|
|
|
|
|
|
def _carrier(height: int, width: int) -> np.ndarray:
|
|
yy, xx = np.mgrid[:height, :width]
|
|
channels = [
|
|
np.cos(2 * np.pi * (14 * xx / width + 14 * yy / height)),
|
|
1.3 * np.cos(2 * np.pi * (14 * xx / width + 14 * yy / height) + 0.2),
|
|
0.8 * np.cos(2 * np.pi * (14 * xx / width + 14 * yy / height) - 0.3),
|
|
]
|
|
return np.stack(channels, axis=2)
|
|
|
|
|
|
def _write_pair(tmp_path: Path, name: str, shape: tuple[int, int], base: int) -> tuple[Path, Path]:
|
|
height, width = shape
|
|
yy, xx = np.mgrid[:height, :width]
|
|
scene = base + 8 * np.sin(2 * np.pi * xx / width) + 5 * np.cos(2 * np.pi * yy / height)
|
|
clean = np.stack([scene, scene + 3, scene - 3], axis=2)
|
|
marked = clean + _carrier(height, width)
|
|
clean_path = tmp_path / f"{name}-clean.png"
|
|
marked_path = tmp_path / f"{name}-marked.png"
|
|
Image.fromarray(np.clip(np.rint(clean), 0, 255).astype(np.uint8)).save(clean_path)
|
|
Image.fromarray(np.clip(np.rint(marked), 0, 255).astype(np.uint8)).save(marked_path)
|
|
return clean_path, marked_path
|
|
|
|
|
|
def test_pair_residual_preserves_float_signal_across_shapes(tmp_path: Path):
|
|
first = _write_pair(tmp_path, "first", (96, 96), 100)
|
|
second = _write_pair(tmp_path, "second", (128, 80), 140)
|
|
|
|
first_residual, first_measurement = probe.pair_residual(*first, size=64)
|
|
second_residual, second_measurement = probe.pair_residual(*second, size=64)
|
|
|
|
assert first_residual.shape == (64, 64, 3)
|
|
assert first_measurement.changed_pixel_fraction > 0.2
|
|
assert second_measurement.width == 80
|
|
assert min(probe.channel_ncc(first_residual, second_residual)) > 0.75
|
|
|
|
|
|
def test_selected_peaks_find_injected_frequency(tmp_path: Path):
|
|
pair = _write_pair(tmp_path, "pair", (96, 96), 100)
|
|
residual, _ = probe.pair_residual(*pair, size=64)
|
|
template = probe.build_template([residual])
|
|
|
|
peaks = probe.select_peaks(template, count=4)
|
|
|
|
assert any(abs(int(dy)) == 14 and abs(int(dx)) == 14 for dy, dx in peaks)
|
|
|
|
|
|
def test_marked_image_scores_above_clean_image(tmp_path: Path):
|
|
pair = _write_pair(tmp_path, "pair", (128, 128), 100)
|
|
residual, _ = probe.pair_residual(*pair, size=128)
|
|
template = probe.build_template([residual])
|
|
peaks = probe.select_peaks(template, count=8)
|
|
|
|
clean_score = probe.score_image(pair[0], template, peaks)
|
|
marked_score = probe.score_image(pair[1], template, peaks)
|
|
|
|
assert marked_score.phase_weighted > clean_score.phase_weighted
|
|
assert marked_score.top_two_channel_phase_12 > clean_score.top_two_channel_phase_12
|
|
|
|
|
|
def test_template_round_trip_does_not_use_pickle(tmp_path: Path):
|
|
template = np.zeros((64, 64, 3), dtype=np.float64)
|
|
template[1, 1] = (1.0, 2.0, 3.0)
|
|
peaks = np.asarray([[1, 1], [2, 2]], dtype=np.int32)
|
|
path = tmp_path / "template.npz"
|
|
|
|
probe.save_template(path, template, peaks)
|
|
loaded_template, loaded_peaks = probe.load_template(path)
|
|
|
|
assert np.array_equal(loaded_template, template)
|
|
assert np.array_equal(loaded_peaks, peaks)
|
|
|
|
|
|
def test_discovery_report_contains_cross_pair_ncc(tmp_path: Path):
|
|
first = _write_pair(tmp_path, "first", (96, 96), 100)
|
|
second = _write_pair(tmp_path, "second", (128, 80), 140)
|
|
first_residual, first_measurement = probe.pair_residual(*first, size=64)
|
|
second_residual, second_measurement = probe.pair_residual(*second, size=64)
|
|
template = probe.build_template([first_residual, second_residual])
|
|
peaks = probe.select_peaks(template, count=4)
|
|
|
|
report = probe.discovery_report(
|
|
[first_residual, second_residual],
|
|
[first_measurement, second_measurement],
|
|
peaks,
|
|
)
|
|
|
|
assert report["pair_count"] == 2
|
|
assert len(report["pairwise"]) == 1
|
|
assert report["wavelet"]["wavelet"] == "db2"
|
|
assert len(report["wavelet"]["bands"]) == 9
|
|
assert report["spectral"]["peaks"]
|
|
assert report["spectral"]["cepstral_peaks"]
|
|
assert report["permutation_control"] is not None
|
|
|
|
|
|
def test_wavelet_report_finds_repeatable_multiscale_carrier(tmp_path: Path):
|
|
first = _write_pair(tmp_path, "first", (96, 96), 100)
|
|
second = _write_pair(tmp_path, "second", (128, 80), 140)
|
|
first_residual, _ = probe.pair_residual(*first, size=64)
|
|
second_residual, _ = probe.pair_residual(*second, size=64)
|
|
|
|
report = probe.wavelet_report([first_residual, second_residual])
|
|
|
|
assert len(report["bands"]) == 9
|
|
assert max(max(band["coefficient_ncc"]) for band in report["bands"]) > 0.7
|
|
assert all(len(band["median_rms"]) == 3 for band in report["bands"])
|
|
|
|
|
|
def test_streaming_pairwise_ncc_matches_explicit_pairs():
|
|
rng = np.random.default_rng(7)
|
|
arrays = [rng.normal(size=(16, 16, 3)) for _ in range(5)]
|
|
normalized = [probe.normalized_channels(values) for values in arrays]
|
|
|
|
explicit = np.mean(
|
|
[probe.channel_ncc(arrays[first], arrays[second]) for first, second in combinations(range(len(arrays)), 2)],
|
|
axis=0,
|
|
)
|
|
squared_norm_sum = np.sum([np.sum(np.square(values), axis=(0, 1)) for values in normalized], axis=0)
|
|
streamed = probe.pairwise_ncc_from_normalized_sum(
|
|
np.sum(normalized, axis=0),
|
|
squared_norm_sum,
|
|
len(normalized),
|
|
)
|
|
|
|
assert np.allclose(streamed, explicit)
|
|
|
|
|
|
def test_streaming_pairwise_ncc_handles_zero_norm_channels():
|
|
arrays = [np.zeros((8, 8, 3)), np.zeros((8, 8, 3))]
|
|
normalized = [probe.normalized_channels(values) for values in arrays]
|
|
|
|
streamed = probe.pairwise_ncc_from_normalized_sum(
|
|
np.sum(normalized, axis=0),
|
|
np.zeros(3),
|
|
len(normalized),
|
|
)
|
|
|
|
assert streamed == (0.0, 0.0, 0.0)
|
|
|
|
|
|
def test_spectral_report_finds_injected_frequency(tmp_path: Path):
|
|
first = _write_pair(tmp_path, "first", (96, 96), 100)
|
|
second = _write_pair(tmp_path, "second", (128, 80), 140)
|
|
first_residual, _ = probe.pair_residual(*first, size=64)
|
|
second_residual, _ = probe.pair_residual(*second, size=64)
|
|
|
|
report = probe.spectral_report([first_residual, second_residual])
|
|
|
|
assert any(abs(peak["dy"]) == 14 and abs(peak["dx"]) == 14 for peak in report["peaks"])
|
|
assert report["phase_coherence_p95"] >= report["phase_coherence_median"]
|
|
|
|
|
|
def test_permutation_control_breaks_exact_pair_repeatability(tmp_path: Path):
|
|
first = _write_pair(tmp_path, "first", (96, 96), 100)
|
|
second = _write_pair(tmp_path, "second", (128, 80), 140)
|
|
first_residual, first_measurement = probe.pair_residual(*first, size=64)
|
|
second_residual, second_measurement = probe.pair_residual(*second, size=64)
|
|
|
|
report = probe.permutation_control(
|
|
[first_residual, second_residual],
|
|
[first_measurement, second_measurement],
|
|
)
|
|
|
|
assert report is not None
|
|
assert np.mean(report["true_pair_ncc"]) > np.mean(report["mismatched_pair_ncc"])
|
|
assert report["mismatched_median_rms"] > report["true_median_rms"]
|