Image Forensics

How to Use Image Forensics

1. Upload Image: Drag and drop any image (JPEG, PNG, WebP) onto the upload zone. The tool loads it at full resolution for pixel-level analysis.
2. Apply Forensic Filters: Click any filter button — ELA (Error Level Analysis), Edge Detect, Noise Map, Grayscale, Invert, Sharpen, Emboss, or individual RGB/Alpha channels. The filtered view updates instantly on the canvas.
3. Analyze ELA: The ELA filter re-compresses the image at 75% quality and amplifies the difference. Brighter areas in the ELA output indicate regions that were edited at a different compression level — a key indicator of manipulation.
4. Review Statistics: Below the filters, see dimensions, pixel count, megapixels, average/max RGB values, aspect ratio, and file size.
5. Inspect Histogram: The RGB histogram shows color distribution — useful for detecting clipping, unusual tonal patterns, or artificial color manipulation.
6. Download: Click "Download Filtered" to save the current filtered view as a PNG for reports or documentation.

Technical Overview & Use Cases

This tool implements multiple image forensic techniques using Canvas pixel manipulation. Error Level Analysis (ELA) works by re-saving the image at a known JPEG quality (75%) and computing the pixel-by-pixel difference amplified 10× — edited regions show brighter because they have different compression histories. Convolution filters (Edge, Sharpen, Emboss) apply 3×3 kernel matrices via nested loops, detecting boundaries and structural features in the pixel data. The noise map calculates per-pixel color channel variance to reveal synthetic patterns invisible to the naked eye.

Real-world use cases:

• Journalism & fact-checking: Verify whether a viral image has been digitally altered — ELA reveals copy-paste regions, clone-stamp areas, and spliced content that maintain different compression signatures.
• Insurance fraud detection: Identify manipulated damage photos in insurance claims by detecting inconsistent compression artifacts and unnatural edge patterns around alleged damage.
• Academic integrity: Detect image manipulation in research papers, scientific publications, and medical imaging where data integrity is legally and ethically critical.

Privacy & Security Guarantee

This tool is part of the FAK LAB ecosystem, founded by Faizan Ahmad Khan Khichi. All forensic analysis runs 100% in your browser using Canvas API pixel operations. Your images — including evidence photos, confidential documents, or sensitive materials — are never uploaded to any server. ELA re-compression, convolution, and histogram computation all happen in local memory. No image data, analysis results, or pixel values are transmitted anywhere.

Frequently Asked Questions

What does a bright spot in ELA mean?

Bright areas in ELA indicate pixels with different compression histories than their surroundings. When an image is edited and re-saved, the edited regions undergo fewer compression passes than the original. This mismatch produces higher error levels (brighter spots) in those regions. However, ELA is not definitive proof — high-contrast edges, text, and fine details also produce bright ELA output naturally.

Can this tool definitively prove an image is fake?

No single tool can conclusively prove manipulation. This forensic toolkit provides indicators that raise suspicion — ELA anomalies, unusual noise patterns, edge discontinuities. Professional forensic analysis combines multiple techniques with expert interpretation. Use these results as preliminary screening, not as legal proof.

Why does ELA show bright edges on unedited photos?

High-contrast boundaries (text, sharp edges, geometric shapes) naturally produce high error levels because JPEG compression handles transitions poorly. This is expected behavior, not evidence of manipulation. When interpreting ELA, look for uniformly bright regions in areas that should have consistent compression — not just bright edges or text.