Search

Forensic iris recognition, as opposed to live iris recognition, is an emerging research area that leverages the discriminative power of iris biometrics to aid human examiners in their efforts to identify deceased persons. As a machine learning-based technique in a predominantly human-controlled task, forensic recognition serves as "back-up" to human expertise in the task of post-mortem identification. As such, the machine learning model must be (a) interpretable, and (b) post-mortem-specific, to account for changes in decaying eye tissue. In this work, we propose a method that satisfies both requirements, and that approaches the creation of a post-mortem-specific feature extractor in a novel way employing human perception. We first train a deep learning-based feature detector on post-mortem iris images, using annotations of image regions highlighted by humans as salient for their decision making. In effect, the method learns interpretable features directly from humans, rather than purely data-driven features. Second, regional iris codes (again, with human-driven filtering kernels) are used to pair detected iris patches, which are translated into pairwise, patch-based comparison scores. In this way, our method presents human examiners with human-understandable visual cues in order to justify the identification decision and corresponding confidence score. When tested on a dataset of post-mortem iris images collected from 259 deceased subjects, the proposed method places among the three best iris matchers, demonstrating better results than the commercial (non-human-interpretable) VeriEye approach. We propose a unique post-mortem iris recognition method trained with human saliency to give fully-interpretable comparison outcomes for use in the context of forensic examination, achieving state-of-the-art recognition performance.

Iris recognition of living individuals is a mature biometric modality that has been adopted globally from governmental ID programs, border crossing, voter registration and de-duplication, to unlocking mobile phones. On the other hand, the possibility of recognizing deceased subjects with their iris patterns has emerged recently. In this paper, we present an end-to-end deep learning-based method for postmortem iris segmentation and recognition with a special visualization technique intended to support forensic human examiners in their efforts. The proposed postmortem iris segmentation approach outperforms the state of the art and in addition to iris annulus, as in case of classical iris segmentation methods - detects abnormal regions caused by eye decomposition processes, such as furrows or irregular specular highlights present on the drying and wrinkling cornea. The method was trained and validated with data acquired from 171 cadavers, kept in mortuary conditions, and tested on subject-disjoint data acquired from 259 deceased subjects. To our knowledge, this is the largest corpus of data used in postmortem iris recognition research to date. The source code of the proposed method are offered with the paper. The test data will be available through the National Archive of Criminal Justice Data (NACJD) archives.

Iris biometrics involves using the pattern of texture in the iris, the colored region of the eye surrounding the pupil, to verify a person’s identity. The field originated in the work of John Daugman,1, 2 and, in recent years, heightened global and domestic security concerns have lead to an increased interest in iris biometric technology in academia, industry, and government.3 The field promises highly accurate identity verification, but a more flexible user interface is required. However, this could lead to less well-controlled image acquisition, so there is also a need for new techniques to handle a broader range of image quality. To this end, we have been supporting the US Government’s Iris Challenge Evaluation programs4 and the current Multiple Biometric Grand Challenge program,5 as well as conducting our own basic research on iris biometrics.6–8 Many factors complicate the use of iris biometrics, such as differences in pupil dilation, the presence of contact lenses, and the eye’s natural aging. Figure 1 shows an example of the type of image used. Current systems largely ignore differences in pupil dilation. They also assume that the natural aging of the eye has no effect on accuracy. Commercial technology can handle some problems that arise with contact lenses, for example, but our research suggests that there are more common, subtle, and smaller effects that are currently ignored. Iris biometrics works best when the pupil is not strongly dilated, and when the degree of dilation is similar in the image used to enroll a person in the system and the image used for verification. The conventional approach ignores differences in pupil dilation and simply converts a region from the iris image into a standard size, and then creates a corresponding binary ‘iris code‘ based on the texture pattern. In current systems of this type, pupil dilation information is discarded. Figure 1. Example iris biometrics image. The eyelids occlude the upper and lower parts of the iris. A shadow is cast on the lowest visible portion, and a specular reflection is seen at the top center of the visible portion. All of these factors complicate iris image analysis.

This book provides a detailed assessment of the state of the art in automated techniques for the analysis of digital mammogram images. Topics covered include a variety of approaches for image processing and pattern recognition aimed at assisting the physician in the task of detecting tumours from evidence in mammogram images. The chapters are written by experts in the field and are revised versions of papers selected from those presented at the "First International Workshop on Mammogram Image Analysis" held in San Jose as part of the 1993 Biomedical Image Processing conference.

Over the past few years, a number of researchers have presented algorithms for computing the aspect graph representation for polyhedra and curved-surface objects. However, currently it is computer from the theoretical standpoint of perfect resolution in the viewpoint, the projected image and the object shape. This means that the aspect graph may include details that an observe could never see in practice. This paper reviews a complete implementation of an algorithm to compute the exact aspect graph of solids of revolution under the perspective projection in 3D space. The we explore the notion of introducing scale into the qualitative aspect graph framework, this providing a mechanism for selecting a level of detail that is large enough to merit explicit representation. Several alternative interpretations of the scale space aspect graph are examined in response to the results produced for an example object by the implemented system.

The Schubmehl-Prein Competition is open each year to high school juniors. The firstplace prize is $1,000, the second-place prize is $500, and the third-place prize is $250. Winning entries have also traditionally been published in the Association for Computing Machinery's Computers and Society online magazine.

This seminar/workshop on ethics and the social impact in computer science, supported by studies of the applicability of alternative teaching and learning strategies, is targeted towards doctoral candidates in computer science whose life-goal is to teach in a university or college setting. Based on the concept of “ethics across the curriculum” the seminar/workshop will prepare future faculty to incorporate ethical and social impact concerns in their technical courses. At the same time they will be exposed to modern teaching and learning techniques that will assist them in making a good start in their teaching careers.

We present an algorithm for recovering a CAD model from a time sequence of range images of an articulated object. Each recovered part is described using an augmented boundary representation which includes the actually observed surfaces of the object part, implied surfaces between it and other connected parts, and occlusion limits to the unseen portion of the part. The algorithm first develops feature correspondences across the frames of the input sequence. Then, after potential connections are inferred between parts based on relative motion, the boundary of each part is computed. This is followed by the calculation of the parameters of validated connections, including any potential causal relations. >

Computed tomographic (CT)-guided stereotaxic procedure have become established in a few major centers but a single optimal system has not yet emerged. At Duke University Medical Center, availability of a stereotaxic frame with design features that lend themselves to CT adaptation made modification of this unit, rather than construction of a new dedicated frame, cost-effective. As other centers may seek to modify available stereotaxic equipment in a similar way, this report documents the modifications ans the early clinical experience in four patients in whom CT-guided stereotaxic biopsy procedures were performed.

Workshops on the theme of “Teaching Ethics and Computing” were sponsored by the National Science Foundation's Undergraduate Faculty Enhancement program. This paper outlines the results of the workshops, available at marathon. csee.usf.edu/~kwb/naf-ufe/. Preparation for the workshops included a survey of videos that are potentially useful in teaching ethics and computing. This paper reviews some of the “best of” these videos.