Abaza A.,Advanced Technologies Group |
Ross A.,West Virginia University |
Hebert C.,Advanced Technologies Group |
Harrison M.A.F.,Advanced Technologies Group |
Nixon M.S.,University of Southampton
ACM Computing Surveys | Year: 2013
Recognizing people by their ear has recently received significant attention in the literature. Several reasons account for this trend: first, ear recognition does not suffer from some problems associated with other noncontact biometrics, such as face recognition; second, it is the most promising candidate for combination with the face in the context of multi-pose face recognition; and third, the ear can be used for human recognition in surveillance videos where the face may be occluded completely or in part. Further, the ear appears to degrade little with age. Even though current ear detection and recognition systems have reached a certain level of maturity, their success is limited to controlled indoor conditions. In addition to variation in illumination, other open research problems include hair occlusion, earprint forensics, ear symmetry, ear classification, and ear individuality. This article provides a detailed survey of research conducted in ear detection and recognition. It provides an up-to-date review of the existing literature revealing the current state-of-art for not only those who are working in this area but also for those who might exploit this new approach. Furthermore, it offers insights into some unsolved ear recognition problems as well as ear databases available for researchers. © 2013 ACM.
Pazinato D.V.,University of Campinas |
Stein B.V.,University of Campinas |
De Almeida W.R.,University of Campinas |
De Werneck R.O.,University of Campinas |
And 6 more authors.
IEEE Journal of Biomedical and Health Informatics | Year: 2016
Background: Pixel-level tissue classification for ultrasound images, commonly applied to carotid images, is usually based on defining thresholds for the isolated pixel values. Ranges of pixel values are defined for the classification of each tissue. The classification of pixels is then used to determine the carotid plaque composition and, consequently, to determine the risk of diseases (e.g., strokes) and whether or not a surgery is necessary. The use of threshold-based methods dates from the early 2000s but it is still widely used for virtual histology. Methodology/Principal Findings: We propose the use of descriptors that take into account information about a neighborhood of a pixel when classifying it. We evaluated experimentally different descriptors (statistical moments, texture-based, gradient-based, local binary patterns, etc.) on a dataset of five types of tissues: blood, lipids, muscle, fibrous, and calcium. The pipeline of the proposed classification method is based on image normalization, multiscale feature extraction, including the proposal of a new descriptor, and machine learning classification. We have also analyzed the correlation between the proposed pixel classification method in the ultrasound images and the real histology with the aid of medical specialists. Conclusions/Significance: The classification accuracy obtained by the proposed method with the novel descriptor in the ultrasound tissue images (around 73%) is significantly above the accuracy of the state-of-the-art threshold-based methods (around 54%). The results are validated by statistical tests. The correlation between the virtual and real histology confirms the quality of the proposed approach showing it is a robust ally for the virtual histology in ultrasound images. © 2014 IEEE.
Shalf J.,Advanced Technologies Group
Scientific Computing | Year: 2010
International Supercomputing Conference 2010 was held in Hamburg, Germany with a focus on measuring the energy efficiency of supercomputing centers. The approach needs metrics that provide a better reflection of real-world use of these machines to run codes and applications. The efficiency at which a system integrates with the overall facility is also very important and can't be overlooked. Academic institutions, laboratories and other HPC end-users have a much more detailed understanding of their HPC application performance and requirements. The BoF session enabled the universities, DOE laboratories and industry to lay out their respective roles in creating this new approach to energy efficiency. Organizations have to look at the end-to-end burden and not shift the energy costs from the systems to the building infrastructure.
Yellampalle B.,Advanced Technologies Group |
McCormick W.B.,Advanced Technologies Group |
Wu H.-S.,Advanced Technologies Group |
Sluch M.,Advanced Technologies Group |
And 3 more authors.
Proceedings of SPIE - The International Society for Optical Engineering | Year: 2014
A key challenge for standoff explosive sensors is to distinguish explosives, with high confidence, from a myriad of unknown background materials that may have interfering spectral peaks. To meet this challenge a sensor needs to exhibit high specificity and high sensitivity in detection at low signal-to-noise ratio levels. We had proposed a Dual-Excitation- Wavelength Resonance-Raman Detector (DEWRRED) to address this need. In our previous work, we discussed various components designed at WVHTCF for a DEWRRED sensor. In this work, we show a completely assembled laboratory prototype of a DEWRRED sensor and utilize it to detect explosives from two standoff distances. The sensor system includes two novel, compact CW deep-Ultraviolet (DUV) lasers, a compact dual-band high throughput DUV spectrometer, and a highly-sensitive detection algorithm. We choose DUV excitation because Raman intensities from explosive traces are enhanced and fluorescence and solar background are not present. The DEWRRED technique exploits the excitation wavelength dependence of Raman signal strength, arising from complex interplay of resonant enhancement, self-absorption and laser penetration depth. We show measurements from >10 explosives/pre-cursor materials at different standoff distances. The sensor showed high sensitivity in explosive detection even when the signalto- noise ratio was close to one (~1.6). We measured receiver-operating-characteristics, which show a clear benefit in using the dual-excitation-wavelength technique as compared to a single-excitation-wavelength technique. Our measurements also show improved specificity using the amplitude variation information in the dual-excitation spectra.© 2014 SPIE.
News Article | December 14, 2016
After the successful launch of UberX in Pittsburgh last September, the company has also begun deploying the driverless vehicles in San Francisco. As is the case in the Steel City, the recent roll out would also have any passengers in the area hail the self-driving vehicle through Uber's mobile app. Unlike the fleet of autonomous Fords in Pittsburgh, however, San Francisco streets will be prowled by the Volvo XC90 UberX version. According to Tech Crunch, there is a difference between these two autonomous cars. The former has been acquired stock off the line whereas Volvo has its proprietary sensor array, which complements Uber's sensors and supercomputers. "The car is one of the reasons we're really excited about this partnership, it's a really tremendous vehicle," Matt Sweeney, head of product at Uber's Advanced Technologies Group, said. "It's Volvo's new SPA, the scalable platform architecture — the first car on their brand new, built from the ground up vehicle architecture, so you get all new mechanical, all new electrical, all new compute." Passengers eager to try the driverless UberX should not be surprised to find that it would pick-them up still manned by a pair of technicians. This includes a safety driver, which could takeover driving when needed and an Uber test engineer to monitor the UberX performance. The inclusion of the engineer is particularly notable at this point. It highlights the fact that the Uber driverless technology is still a work in progress, learning from experience to develop technologies and new measures. The outcome of the constant monitoring can be demonstrated in the way the Volvo UberX is outfitted with fewer sensors than the second-generation driverless Fords. Insights gained in the Pittsburgh pilot revealed that some sensors are no longer needed, so they have been eliminated in the latest batch of UberX. Passengers in San Francisco should take heart, however, that despite the streamlined technology, the Volvo XC90 is still outfitted with the complete array of technologies necessary to guarantee safe and successful driverless experience. This includes the traditional optical cameras and radar as well as the LiDar and ultrasonic sensors. Uber has also been testing UberX in San Francisco for some time. The sheer length of this testing period has been largely attributed to the goal of learning the San Francisco streetscape, which is significantly different from the urban topography of Pittsburgh. "With its challenging roads and often varied weather, Pittsburgh provided a wide array of experiences," Uber said in a press statement. "San Francisco comes with its own nuances including more bikes on the road, high traffic density and narrow lanes." There is an expectation, therefore, that UberX should already know the urban landscape like the back of its, well, hand. © 2017 Tech Times, All rights reserved. Do not reproduce without permission.