Automatic tattoo matching technique for suspect identity

Abstract EN

At the few years ago, the attention of drawing tattoo on a part of the body is more popular, especially in younger populations with age between 20 and 30 years old.

Lawenforcement agencies tend to use the tattoo images as a biometric for the identificationof a suspect or victim and routinely catalog tattoo images for this process.

So the need to development an automatic system for the verification process and matching tattoo sketch (that obtained from the description of the eyewitness) with a stored tattoo image became most important.

The attempts in this scope are limited because of the challenges in matching sketch to image.

In this work a set of features that are invariant against changes in (scale, change, mirroring, rotation and translation) are used.

Moment Invariants features are extracted from tattoo images and combined with the eigenvalues to construct feature vector with 13 values.

These vectors are used in the process of matching between sketch tattoo and image tattoo.

A total number of images that used here is 120 for sketch tattoos and tattoo images, and the obtained results were 85 % of images correctly matched its sketch features