Tamper Detection
Tamper detection encompasses the techniques used to identify unauthorised modifications made to a digital or physical document. It analyses visual, structural, and digital anomalies to determine whether a document has been altered.
Tamper detection plays a central role in the fight against document fraud. While document authentication verifies whether a document is genuine overall, tamper detection focuses specifically on identifying modifications made to an originally authentic document: photo replacement, text alteration, changes to dates or amounts.
Detection techniques analyse multiple levels: at the pixel level, searching for compression inconsistencies (a pasted element will have a different JPEG compression level from the rest of the image), cloning (copy-pasting areas to conceal text), and interpolation (resizing leaves detectable traces). At the structural level, analysis covers font consistency, text alignment, margins, and the document's layout grid.
AI-based solutions go beyond visual analysis by exploiting file metadata (modification history, software used, timestamps), sensor noise analysis (each camera leaves a unique fingerprint), and detection of statistical anomalies in colour and texture distribution. This multi-layered approach enables the detection of even highly sophisticated tampering.
Regulations
Real-world examples
- 1.A verification system detects that a payslip has been modified: analysis reveals the net salary figure was digitally altered, with JPEG compression artefacts different from the rest of the document.
- 2.Analysis of a tax notice submitted for a mortgage application reveals the reference taxable income has been changed: the figures display a slightly different font and misaligned positioning compared to the rest of the document.
- 3.A bank statement presented as part of a rental application shows signs of tampering: clone detection analysis reveals that certain credit lines were duplicated to inflate the apparent account balance.