Clone Detection Scanner

Detect copy-move forgery by finding duplicated regions within an image. Uses DCT fingerprinting to match visually identical blocks in different locations.

500+ images scanned

Drag & drop a file here, or

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

DCT Fingerprinting

Extracts frequency-domain fingerprints from overlapping image blocks to find visually identical regions.

Region Clustering

Groups matched blocks into coherent clone regions using spatial clustering with color-coded visualization.

100% Private

All analysis runs locally in your browser via Web Worker. Nothing is uploaded to any server.

How clone detection works

Copy-move forgery is one of the most common forms of image manipulation — the forger copies a region of the image and pastes it over an object to conceal or duplicate it. Clone detection finds these duplicated regions by dividing the image into overlapping 16×16 blocks, computing a DCT (Discrete Cosine Transform) fingerprint for each block, then sorting and comparing fingerprints to find matches that are spatially separated. Matched blocks are clustered into coherent regions and highlighted on the clone map. For compression-based splice detection, use the JPEG Ghost Scanner. For pixel-level error analysis, see the ELA Scanner.

Sensitivity and false positives

The sensitivity slider (1–10) controls how closely blocks must match to be flagged as clones. At low sensitivity (1–3), only near-identical blocks are detected — useful for reducing noise on textured images. At high sensitivity (7–10), more subtle matches are caught, but natural repetitive textures like grass, sky gradients, brick walls, or fabric patterns may trigger false positives. Start at the default (5) and adjust based on what the image contains. If you see many highlighted regions on a clearly unedited photo, lower the sensitivity. If you suspect subtle cloning, increase it.

Who uses clone detection

Journalists and fact-checkers use clone detection to verify whether objects have been duplicated or removed from news photos. Insurance investigators check claim images for duplicated damage or removed evidence. Legal teams examine photographic evidence for signs of tampering. OSINT researchers analyze satellite imagery and social media photos for copy-paste manipulation. Academic integrity reviewers check research images in scientific publications. For a full forensic workflow, combine with the Authenticity Checker and the Thumbnail Scanner.

Clone detection vs ELA

ELA reveals areas with inconsistent compression history — it catches edits from any source (pasting from another image, retouching, AI inpainting). Clone detection specifically finds regions duplicated within the same image. ELA shows you where something changed; clone detection shows you what was copied. These tools are complementary — an object removed by clone stamping will show up in both analyses. For checking whether the image is AI-generated entirely, add the AI Detector. For hidden data that might reveal editing history, try the Stego Scanner.

Copy-move forgery detection is a cornerstone of image forensics research. The DCT fingerprinting approach used here is robust against minor JPEG recompression and small color adjustments, though geometric transformations (rotation, scaling) of the cloned region may reduce detection accuracy. For maximum forensic coverage, pair clone detection with the photo forensics hub which links to all available analysis tools. All clone detection processing runs entirely in your browser via a Web Worker — your images are never uploaded to any server. For batch analysis of multiple photos, see the Batch Scanner. Learn more in our articles on detecting photo edits and fake photo detection methods.

Frequently Asked Questions

What is copy-move forgery?

Copy-move forgery is when someone copies a region of an image and pastes it in a different location within the same image, typically to hide an object or duplicate something. Clone detection finds these duplicated regions by comparing visual fingerprints of overlapping blocks across the entire image.

How does DCT fingerprinting work?

The image is divided into overlapping 16x16 pixel blocks. Each block is converted to grayscale, pooled to 8x8, then transformed using the Discrete Cosine Transform (DCT) to extract frequency-domain features. The top 16 DCT coefficients form a compact fingerprint. Blocks with matching fingerprints that are spatially separated indicate potential cloning.

Why am I getting false positives on textures?

Natural repetitive patterns like grass, brick walls, fabric, sky gradients, and tiled surfaces contain genuinely similar visual blocks. The scanner cannot distinguish between natural repetition and intentional cloning. Lower the sensitivity slider to reduce these false matches, and focus on unexpected matches in non-repetitive areas of the image.

Can this detect rotated or scaled clones?

The current implementation detects exact copy-paste and minor JPEG recompression but has limited robustness against geometric transformations like rotation, scaling, or mirroring of the cloned region. For manipulations involving geometric changes, complement this tool with ELA Scanner which analyzes compression patterns regardless of geometry.

Is my image uploaded to a server?

No. All clone detection processing runs entirely in your browser using a Web Worker. The image never leaves your device. No data is transmitted, stored, or logged.

What do the colored regions on the clone map mean?

Each numbered colored box represents a cluster of blocks with matching fingerprints. Different colors distinguish separate clone regions. White lines connect individual matched block pairs between their source and destination locations. Multiple regions in different colors suggest multiple copy-paste operations.