Tutorial-1: New criminal and victim identification methods for sexual offenses against women and children

With recent advances in Internet and multimedia technology, the involvement of digital images in sexual offenses against women and children has been increasing significantly. These sexual offenses can be rapes, child sexual abuses and gang rapes. Even in one child sexual abuse case, evidence images and videos can be over 100 gigabytes. In Canada alone, Cybertip.ca received over 21,000 tips about online child exploitation between 2002 and 2008 and found 12,696 websites offering child sexual abuse images. Identifying criminals from this digital evidence can be very challenging because they are usually careful to hide or cover their faces and tattoos. All traditional biometric identification methods, e.g., faces, tattoos and fingerprints are not applicable to these evidence images. This talk will present a set of new computational methods for identifying criminals and victims in this digital evidence.