With continual advances in Internet capability, in addition to its global and decentralized nature, the Internet along with different social networking sites are experiencing a boom in demand and supply. Recent study found that the social networking sites like Facebook, Twitter, and MySpace are providing a forum for paedophiles to share child pornography. With the advent of sophisticated digital technology, Law Enforcement Agency (LEAs) around the world dealing with child pornography facing real challenge to combat with the technologically-savvy paedophiles. The major challenge in child pornography lies in authentic detection of children face in an image. The main objective of this research is to present a novel framework for a dedicated child face detection tool, where we will use child’s face specific contextual contexts and visual cues that are based on new knowledge in terms of features or contexts representatives of child’s skin and face. The proposed technique can estimate age categorically – adult or child based on a new hybrid feature descriptor, called Luminance Invariant & Geometrical Relation based Descriptor (LIGRD). LIGRD is composed of some low and high-level features, which are found to be effective in characterizing the local appearance in terms of chromaticity, texture, and geometric relational information of few facial visual cues simultaneously. Comparison of our experimental results with that of another recently published work reveals our proposed approach yields the highest precision and recall, and overall accuracy in recognition.
|Title of host publication||Deep Learning Applications for Cyber Security|
|Editors||Mamoun Alazab, MingJian Tang|
|Number of pages||26|
|Publication status||Published - 2019|
|Name||Advanced Sciences and Technologies for Security Applications|
|Publisher||Springer International Publishing|