Chris Bensing Boehnen

A multi-sample standoff multimodal biometric system

The data captured by existing standoff biometric systems typically has lower biometric recognition performance than their close range counterparts due to imaging challenges, pose challenges, and other factors. To assist in overcoming these limitations systems typically perform in a multi-modal capacity such as Honeywells Combined Face and Iris (CFAIRS) [21] system. While this improves the systems performance, standoff systems have yet to be proven as accurate as their close range equivalents. We will present a standoff system capable of operating up to 7 meters in range. Unlike many systems such as the CFAIRS our system captures high quality 12 MP video allowing for a multi-sample as well as multimodal comparison. We found that for standoff systems multi-sample improved performance more than multimodal. For a small test group of 50 subjects we were able to achieve 100% rank one recognition performance with our system on standoff recognition of noncooperative subjects.

Gaze estimation for off-angle iris recognition based on the biometric eye model

Iris recognition is among the highest accuracy biometrics. However, its accuracy relies on controlled high quality capture data and is negatively affected by several factors such as angle, occlusion, and dilation. Non-ideal iris recognition is a new research focus in biometrics. In this paper, we present a gaze estimation method designed for use in an off-angle iris recognition framework based on the ORNL biometric eye model. Gaze estimation is an important prerequisite step to correct an off-angle iris images. To achieve the accurate frontal reconstruction of an off-angle iris image, we first need to estimate the eye gaze direction from elliptical features of an iris image. Typically additional information such as well-controlled light sources, head mounted equipment, and multiple cameras are not available. Our approach utilizes only the iris and pupil boundary segmentation allowing it to be applicable to all iris capture hardware. We compare the boundaries with a look-up-table generated by using our biologically inspired biometric eye model and find the closest feature point in the look-up-table to estimate the gaze. Based on the results from real images, the proposed method shows effectiveness in gaze estimation accuracy for our biometric eye model with an average error of approximately 3.5 degrees over a 50 degree range.

Off-angle iris correction using a biological model

This work implements an eye model to simulate corneal refraction effects. Using this model, ray tracing is performed to calculate transforms to remove refractive effects in off-angle iris images when reprojected to a frontal view. The correction process is used as a preprocessing step for off-angle iris images for input to a commercial matcher. With this method, a match score distribution mean improvement of 11.65% for 30 degree images, 44.94% for 40 degree images, and 146.1% improvement for 50 degree images is observed versus match score distributions with unmodified images.

Facial Analytics: From Big Data to Law Enforcement

Facial analytics is an emerging soft-biometric technology that examiners can use to contextualize images of people without encroaching on their privacy.

An Iris Segmentation Algorithm based on Edge Orientation for Off-angle Iris Recognition

Iris recognition is known as one of the most accurate and reliable biometrics. However, the accuracy of iris recognition systems depends on the quality of data capture and is negatively affected by several factors such as angle, occlusion, and dilation. In this paper, we present a segmentation algorithm for off-angle iris images that uses edge detection, edge elimination, edge classification, and ellipse fitting techniques. In our approach, we first detect all candidate edges in the iris image by using the canny edge detector; this collection contains edges from the iris and pupil boundaries as well as eyelash, eyelids, iris texture etc. Edge orientation is used to eliminate the edges that cannot be part of the iris or pupil. Then, we classify the remaining edge points into two sets as pupil edges and iris edges. Finally, we randomly generate subsets of iris and pupil edge points, fit ellipses for each subset, select ellipses with similar parameters, and average to form the resultant ellipses. Based on the results from real experiments, the proposed method shows effectiveness in segmentation for off-angle iris images.

Impact of environmental factors on biometric matching during human decomposition

Automatic recognition systems are valuable tools for identifying unknown deceased individuals. Immediately after death, fingerprint and face biometric samples are easy to collect using standard sensors and can be easily matched to antemortem biometric samples. Even though early postmortem fingerprints and facial images have been used for identification purposes for decades, there are no studies that track these biometrics through the later stages of decomposition to determine the length of time they remain viable. This paper discusses a multimodal dataset of finger-prints, faces, and irises from twelve donated human subjects that decomposed outdoors under natural conditions. Results include predictive models relating time and temperature, measured as Accumulated Degree Days (ADD), and season (winter, spring, summer), to the probability of automatic verification using a commercial algorithm.

3D face analysis for demographic biometrics

Despite being increasingly easy to acquire, 3D data is rarely used for face-based biometrics applications beyond identification. Recent work in image-based demographic biometrics has enjoyed much success, but these approaches suffer from the well-known limitations of 2D representations, particularly variations in illumination, texture, and pose, as well as a fundamental inability to describe 3D shape. This paper shows that simple 3D shape features in a face-based coordinate system are capable of representing many biometric attributes without problem-specific models or specialized domain knowledge. The same feature vector achieves impressive results for problems as diverse as age estimation, gender classification, and race classification.

Field trial of a highly portable coded aperture gamma ray and 3D imaging system

Given the growing concerns surrounding declared and undeclared nuclear activities around the world, a novel multi-modal imaging system combining gamma imaging, visible-light imaging, and multi-resolution 3D imaging has been developed at the Oak Ridge National Laboratory (ORNL). As part of an international collaborative effort, this research aims at providing a portable imaging system to assess, monitor and/or detect nuclear material dispersion in nuclear facilities. This novel imaging system includes (1) a coded-aperture gamma-ray imager that provides a map of all major radiological sources present in the field-of-view and (2) a low-resolution stereo system to retrieve high-level 3D scene information. Using a unique approach to combine the data a 3D model of the scene can be rendered with the gamma information as a projected texture, therefore allowing easy visualization of the location of all detected radiological sources. In this paper, we give a summary of this research by first presenting the instruments, then by detailing our approach to project gamma information on the high-resolution 3D point-cloud, and presenting the results of the first field trial of the system.

A standoff multimodal biometric system

Biometric authentication modalities such as face and iris recognition provide a minimally invasive means to uniquely identify individuals and verify identity. Most commercially available biometric systems perform authentication based on a single modality and/or only capture features from a short distance. Most existing standoff iris sensors capture at a lower frame rate, lower resolution, and wider field of view than our system. This work describes the design of a prototype standoff biometric sensor. The complete system is comprised of three sensors, their respective software control modules, and a command and control graphical user interface. Sensor 1 is a high resolution monochrome camera with telephoto zoom and video rate image acquisition. Co-aligned, collimated near infrared (NIR) light emitting diodes (LEDs) provide controlled illumination to facilitate the capture of face and iris images. Sensor 2 is a monochrome stereo camera that acquires low resolution frontal face images and scene depth information. Sensor 3 is dual spectrum camera that acquires pixel registered visible and NIR images at video rate. Ambient room light and NIR flood LEDs provide illumination for capturing profile face and gait images. Real-time analysis of the stereo camera output provides feedback for pan, tilt, zoom, and focus of the sensor platform. A Modularized software control system provides scalability and flexible management. Commodity hardware can be used to control all system components with the exception of sensor 1.

Off-Angle Iris Correction Methods

In many real-world iris recognition systems, obtaining consistent frontal images is problematic do to inexperienced or uncooperative users, untrained operators, or distracting environments. As a result many collected images are unusable by modern iris matchers. In this chapter, we present four methods for correcting off-angle iris images to appear frontal which makes them compatible with existing iris matchers. The methods include an affine correction, a retraced model of the human eye, measured displacements, and a genetic algorithm optimized correction. The affine correction represents a simple way to create an iris image that appears frontal but it does not account for refractive distortions of the cornea. The other method account for refraction. The retraced model simulates the optical properties of the cornea. The other two methods are data-driven. The first uses optical flow to measure the displacements of the iris texture when compared to frontal images of the same subject. The second uses a genetic algorithm to learn a mapping that optimizes the Hamming Distance scores between off-angle and frontal images. In this paper, we hypothesize that the biological model presented in our earlier work does not adequately account for all variations in eye anatomy and therefore the two data-driven approaches should yield better performance. Results are presented using the commercial VeriEye matcher that show that the genetic algorithm method clearly improves over prior work and makes iris recognition possible up to 50\(^\circ \) off-angle.