Forgery Heatmap

Combines four forensic analyses — ELA, noise consistency, JPEG ghost, and clone detection — into a single probability heatmap. Regions where multiple methods converge are highlighted as suspicious.

800+ images scanned for forgery

Drag & drop a file here, or

JPG PNG WebP GIF BMP AVIF Max 50MB • Processed in browser

4-in-1 Analysis

Runs ELA, noise consistency, JPEG ghost detection, and clone detection simultaneously. Each method catches different manipulation types.

Probability Heatmap

Results are combined into a single color-coded overlay. When multiple methods flag the same region, confidence increases — reducing false positives.

100% Private

All four analyses run locally in your browser via Web Worker. Your image is never uploaded to any server.

Why combine multiple forensic methods

No single forensic analysis catches every type of image manipulation. ELA detects re-compression artifacts from cut-and-paste edits. Noise analysis reveals airbrushing and content-aware fill. JPEG ghost detection identifies double-compressed regions saved at different quality levels. Clone detection finds duplicated areas. By running all four simultaneously and weighting their outputs, the Forgery Heatmap reduces false positives — a region flagged by multiple independent methods is far more likely to be genuinely manipulated. For deep single-method analysis, use the dedicated ELA Scanner or Noise Analysis tools.

How the probability heatmap works

The image is divided into 16×16 pixel blocks. Each block is scored independently by all four analysis methods, producing a probability value between 0 and 1. These per-method scores are then combined using a weighted average — ELA and noise analysis receive higher weight because they are more reliable across image formats, while JPEG ghost and clone detection provide supporting evidence. The final combined probability is mapped to a green-to-red color gradient and overlaid on the original image. Regions where multiple methods converge get brighter colors, while areas with conflicting signals stay cooler. For compression-specific analysis, the JPEG Ghost Scanner provides deeper quality sweep.

When to use the Forgery Heatmap vs individual tools

The Forgery Heatmap is ideal as a first-pass screening tool — upload a suspicious image and immediately see which regions need attention. For investigative deep-dives, switch to individual tools: Clone Detection for copy-paste forensics with visual match lines, JPEG Structure Analyzer for quantization table fingerprinting, or the Authenticity Checker for a comprehensive multi-signal assessment including C2PA metadata. The Forgery Heatmap trades depth for breadth — it runs faster than four separate analyses and highlights where to look next.

Limitations and false positives

No automated tool can definitively prove or disprove image manipulation. Certain legitimate image features trigger false positives: hard shadow edges, areas with extreme contrast, heavily textured regions like foliage or fabric, and images that have been resized or re-compressed multiple times. Screenshots and digitally created graphics will often score high because they lack the noise characteristics of camera sensors. Always interpret results in context and combine with visual inspection. For AI-specific detection, see the AI Detector. Explore all available forensic tools for a complete analysis workflow. Learn more in our articles on detecting edited photos and fake photo detection methods.

Frequently Asked Questions

What is a forgery heatmap?

A forgery heatmap is a visual overlay that highlights regions of an image where multiple forensic analysis methods detect signs of manipulation. Each pixel block is color-coded from green (normal) through yellow and orange to red (high probability of tampering). Unlike single-method tools, the heatmap combines four independent analyses to reduce false positives and increase confidence.

Which four analyses does this tool run?

The tool runs Error Level Analysis (ELA) to detect re-compression inconsistencies, noise consistency analysis to find airbrushed or cloned regions, JPEG ghost detection to identify double-compressed areas saved at different quality levels, and clone detection to locate copy-pasted regions. Each method is optimized for speed while maintaining detection accuracy.

How should I interpret the overall score?

The overall score (0-100%) represents the average manipulation probability across all image blocks. Below 15% is consistent with an unmodified image. Between 15-35% suggests minor retouching or natural variations. Between 35-60% indicates possible manipulation worth investigating. Above 60% means strong convergence across multiple methods, suggesting significant editing. Always combine the score with visual inspection of the highlighted regions.

Why does this work better than running one analysis?

Different manipulation techniques leave different traces. A skilled editor might defeat ELA by matching compression quality, but the noise pattern would still be inconsistent. Clone stamping might preserve noise levels but would be caught by the clone detection fingerprinting. By combining methods, the heatmap catches manipulation that would slip past any single analysis. When two or more independent methods flag the same region, confidence in the detection increases significantly.

Are my images uploaded to a server?

No. All four analyses run entirely in your browser using a Web Worker thread. Your images never leave your device. No image data is transmitted, stored, or logged on any server. The tool works offline once the page has loaded.

Can I adjust which analyses are used?

Yes. Each of the four analysis methods can be toggled on or off using the checkboxes in the controls panel. You can also adjust the sensitivity slider to control detection thresholds and the overlay opacity to see more or less of the heatmap over the original image. After changing settings, click Re-run to re-analyze with the new configuration.