Block-Matching
Definition
A copy-move detection strategy that divides the image into overlapping fixed-size blocks, computes a compact feature vector per block, sorts vectors, and identifies suspiciously similar block pairs. The spatial relationship between matched pairs reveals the direction and magnitude of the clone operation.
- Domain
- Copy-move image forgery detection
- Method
- Overlapping fixed-size blocks compared via feature vectors
- Output
- Pairs of visually similar blocks at different image locations
- Extra signal
- Vector between matched pairs shows clone shift direction/magnitude
Common questions
Why are the blocks overlapping rather than a simple non-overlapping grid?+
Overlapping blocks ensure that a cloned region is captured regardless of where its boundary falls relative to the grid; a non-overlapping tiling could split a duplicated area awkwardly and miss the match, while overlap trades extra computation for more reliable detection.
What limits block-matching detection in practice?+
It is computationally expensive on large images because of the number of block comparisons required, and it can miss clones that were rotated, scaled, or heavily post-processed before pasting, since the feature vectors for those blocks no longer match closely with the source region.
Related terms
- Affine Transform Estimation
- The recovery of the geometric transformation (rotation, scaling, shear) relating a copied region to its source, using matched keypoint pairs. Geometrically consistent...
- Chromatic Aberration
- Colour fringing at high-contrast edges caused by differential refraction of light wavelengths by a lens. The pattern is lens-specific and inconsistent across...
- Copy-Move Forgery
- An intra-image manipulation in which a patch copied from one location within the image is pasted over another location in the same...
- Localisation Map
- The output of a forgery detection algorithm that marks, at pixel or block resolution, which regions of the image are identified as...
- Noiseprint
- A CNN-based camera-model fingerprint extractor by Cozzolino and Verdoliva. Applied to deepfakes, it reveals inconsistency between the camera fingerprint in the genuine...
- Passive Forgery Detection
- Detection that operates on the image data alone, without any pre-embedded watermark or signature. Distinguished from active methods such as digital watermarking...
- SIFT (Scale-Invariant Feature Transform)
- A keypoint descriptor algorithm that detects interest points and computes 128-dimensional descriptors invariant to scale and rotation. In forgery detection, matching SIFT...
- Splicing
- An inter-image manipulation in which content from one or more external images is inserted into a target image. Splicing introduces cross-boundary inconsistencies...