Click here for pdf of Dr. Vallone's talk.
Over the last 10 years DNA testing has become increasingly popular for human
identification purposes. DNA typing is commonly used for paternity testing,
identification of mass fatality victims and forensic human identification.
The primary role of DNA typing experiments is to differentiate between
individuals based on their genetic blueprint. Typically, 10 to 15 genetic
markers known to exhibit variation in the human genome are probed. With
current forensic testing methods the probability a random match between 2
unrelated individuals is up to 1 in 3 trillion.
Currently, DNA typing is conducted in approximately 8 to 10 hours. The process includes the extraction of genomic DNA, quantitation, multiplex PCR amplification, and fragment length detection. Currently used commercial multiplex short tandem repeat (STR) typing kits and instrumentation are not optimized or intended for rapid DNA typing.
With the continuing development of miniaturization technologies such as microfluidic and micro-capillary devices, there is a growing interest in developing an integrated DNA typing system capable of going from a buccal swab to a DNA profile in less than 2 hours. Such a rapid DNA typing device could be used, for example, for initial screening at a crime scene, at a border, or at airports. One component of such an integrated platform involves the reduction in time required for multiplex PCR amplification. The potential of reducing the required PCR amplification time may also benefit laboratories typing single-source reference samples.
Further work with various non-standard fast protocols has resulted in decreasing the PCR amplification time to less than 20 minutes for a 16 locus commercial STR typing kit. Results will be shown for STR multiplex assays amplified on various thermal cycling platforms.
NEW - Hear Hany Farid's August 27, 2009 interview on NPR.
Abstract. Photography lost its innocence many years ago. Shortly after the first commercially available camera was introduced, photographs were being manipulated and altered. With the advent of high-resolution digital cameras, powerful personal computers and sophisticated photo-editing software, the manipulation of digital images is becoming more common. We are seeing the impact of these technologies in nearly every corner of our lives. As the technology that allows for digital media to be manipulated and distorted is developing at break-neck speeds, our understanding of the technological, ethical and social implications is lagging behind. I will discuss some of these issues and describe computational techniques which we have developed for detecting tampering in digital media. Operating in the absence of digital watermarks or signatures, these techniques quantify and detect statistical correlations that result from specific forms of digital tampering.
Click here for pdf of Dr. O'Toole's talk.
Abstract. The human face is a complex 3D surface with an overlying reflectance (albedo) map. Psychologists and computer-vision researchers have commonly focused on understanding and quantifying the "unique" identity information in faces in a way that is robust to photometric variation (e.g., illumination, resolution, and viewpoint). More recently, this focus has broadened to consider changes in a person's facial appearance due to natural phenomena such as aging. I will present a set of interconnected studies aimed at understanding the source of the "unique" identity information in faces and bodies. These experiments begin with a look at the role of 3D shape and 2D reflectance information in human face recognition accuracy. A key characteristic of human face representations, long clear from perceptual studies of facial caricatures, is that individual faces are encoded relative to a population "average". Caricature studies offer insight both into the nature of human visual codes for faces and into ways of connecting the representation of an individual face to a population of "non-target" faces. Moving out of the laboratory and into more natural settings, it is clear that in most cases we experience faces and bodies in motion. Dissecting the quality of identity information in dynamic versus static presentations of faces and bodies, and combining useful biometrics, is proving to be a fruitful approach to bridging the gaps inherent in any single biometric. In closing, I will argue that human-machine comparisons, both on a quantitative and qualitative level, can offer new and perhaps more realistic expectations for face and person identity biometrics.