20% Erkennungsgewichtung

Frequenzanalyse

DCT-Spektrumanalyse

Die Frequenzdomänenanalyse untersucht die Verteilung von Hoch- und Niederfrequenzkomponenten in einem Bild. KI-generierte Bilder fehlt typischerweise das natürliche Hochfrequenzrauschen echter Fotografien.

78%
Durchschnittliche Genauigkeit
20%
Erkennungsgewichtung
Frequency Analysis Illustration

Wie es Funktioniert

Mit der Diskreten Kosinustransformation (DCT) zerlegen wir das Bild in seine Frequenzkomponenten. Echte Fotografien enthalten reichhaltige Hochfrequenzdetails von Kamerasensoren, während KI-Bilder tendenziell glatter sind.

Real Photos

  • • Rich high-frequency content
  • • Natural camera sensor noise
  • • Sharp edges and details
  • • High_freq_ratio > 0.20

AI Generated

  • • Smooth, low-frequency dominant
  • • Uniform synthetic patterns
  • • Blurry fine details
  • • High_freq_ratio < 0.15

Technical Implementation

DCT (Discrete Cosine Transform)

We decompose images into frequency components using 2D DCT, then analyze the energy distribution between high and low frequency regions.

Threshold
0.15 (AI likely)
Processing
512x512 max
Library
scipy.fftpack

Frequently Asked Questions

What is frequency analysis for AI detection?

Frequency analysis uses DCT (Discrete Cosine Transform) to decompose images into frequency components. AI images typically lack the high-frequency detail found in real photographs.

Why do AI images lack high frequencies?

AI diffusion models generate images at lower resolutions and upsample them. This process smooths out fine details, reducing high-frequency content compared to sensor-captured photos.

What is the high-frequency ratio threshold?

Real photos typically have a high-frequency ratio above 0.20. AI-generated images often fall below 0.15, indicating their synthetic smooth nature.

How is the DCT applied?

We convert images to grayscale, apply 2D DCT, and separate the spectrum into low and high frequency regions. The ratio of energy in these regions indicates image origin.

Does image sharpening affect results?

Artificial sharpening can boost high frequencies, but the pattern differs from natural sensor noise. Our analysis considers both ratio and distribution characteristics.

Why is frequency analysis weighted at 20%?

Frequency analysis is highly reliable for detecting diffusion models but can be affected by image filters. The 20% weight reflects its strong but not absolute discriminative power.

What's the difference from FFT analysis?

DCT produces only real values (no complex numbers) and has better energy compaction for image signals. It's also what JPEG compression uses, making it well-suited for forensics.

Can upscaled AI images fool this?

AI upscalers add synthetic high-frequency detail, but these patterns are uniform and lack the natural variation of camera-captured fine detail. Detection remains effective.

What image size is optimal?

We process images at 512x512 maximum for analysis. Higher resolution provides more frequency data, but the ratios remain consistent across sizes.

Does this work on all AI generators?

Frequency analysis is particularly effective against diffusion models (DALL-E, Midjourney, Stable Diffusion). Older GAN-based generators show different but still detectable patterns.

Verwandte Methoden

ML-Erkennung

Unsere ML-Erkennung nutzt modernste Transformer-Modelle, die auf Millionen von Bildern trainiert wurden.

PRNU-Analyse

Photo Response Non-Uniformity (PRNU) detects unique camera sensor fingerprints from manufacturing imperfections. AI images cannot replicate these authentic sensor signatures.

Gradientenanalyse

Analyzes edge patterns and texture characteristics using Sobel, Canny, and Laplacian operators. AI images often have unnaturally smooth or uniform gradients.

Rauschmuster

Echte Fotografien enthalten einzigartige Rauschmuster von Kamerasensoren, die über das Bild variieren. KI-generierte Bilder haben eine unnatürlich gleichmäßige Rauschverteilung.

Metadatenanalyse

Bildmetadaten enthalten wertvolle Hinweise auf den Ursprung. Wir analysieren EXIF-Daten, Softwaresignaturen und andere eingebettete Informationen, um KI-Generierungswerkzeuge zu identifizieren.

GAN-Fingerabdruck

Erkennt GAN-spezifische Artefakte wie Schachbrettmuster, Farbbanding und spektrale Anomalien.

Texturanalyse

Local Binary Pattern Analyse für Texturanomalien in KI-generierten Bildern.

Anatomische Erkennung

KI-Bildgeneratoren erzeugen oft anatomische Fehler, die Menschen sofort als falsch erkennen. Wir nutzen Computer Vision, um diese verräterischen Fehler zu erkennen.

C2PA-Verifizierung

C2PA (Coalition for Content Provenance and Authenticity) ist ein Industriestandard zur Verfolgung von Ursprung und Geschichte digitaler Inhalte durch kryptografische Signaturen.

Semantic Inconsistency Detection

Detects logical inconsistencies like incorrect shadows, impossible perspectives, distorted reflections, and violations of physical laws that AI often produces.

Human Biometric Analysis

Uses MediaPipe to analyze human anatomy for incorrect finger counts, asymmetric eyes, unnatural skin texture, and other anatomical anomalies common in AI-generated faces.

Lighting Physics Validation

Validates light source consistency, shadow direction physics, specular highlight accuracy, and color temperature uniformity across the image.

Compression Artifact Analysis

Analyzes JPEG compression artifacts to estimate quality levels and detect re-compression patterns that indicate image manipulation or AI generation.

Edge Sharpness Analysis

Analyzes sharpness distribution across the image and validates depth-of-field consistency. AI often produces unnaturally uniform sharpness.

Statistical Pattern Analysis

Analyzes statistical properties including Shannon entropy, histogram patterns, and Benford's Law compliance to detect synthetic image characteristics.

Chromatic Aberration Analysis

Detects the absence of chromatic aberration (color fringing) that real camera lenses produce. AI images lack these optical artifacts.

Micro-Texture Analysis

Analyzes microscopic texture patterns for repetition, uniformity, and unnatural randomness that AI generators often exhibit.

Color Palette Analysis

Analyzes color distribution including saturation levels, color diversity, and white balance consistency. AI images often have oversaturated colors.

Ihr Bild Prüfen

Alle Methoden werden mit gewichteter Bewertung kombiniert, um ein endgültiges Urteil mit Konfidenzniveau zu erzeugen.

Jetzt Ausprobieren