“To faithfully reproduce a real-world object in the VR/AR environment, we need to replicate the 3D geometry and appearance of the object,” says Min H. Kim, associate professor of computer science at KAIST in South Korea and lead author of the research. “Traditionally, this has been either done manually by 3D artists, which is a labor-intensive task, or by using specialized, expensive hardware. Our method is straightforward, cheaper and efficient, and reproduces realistic 3D objects by just taking photos from a single camera with a built-in flash.”

Kim and his collaborators, Diego Gutierrez, professor of computer science at Universidad de Zaragoza in Spain, and KAIST PhD students Giljoo Nam and Joo Ho Lee, will present this new work at SIGGRAPH Asia 2018 in Tokyo 4 December to 7 December. The annual conference features the most respected technical and creative members in the field of computer graphics and interactive techniques, and showcases leading edge research in science, art, gaming and animation, among other sectors.

Existing approaches for the acquisition of physical objects require specialized hardware setups to achieve geometry and appearance modeling of the desired objects. Those setups might include a 3D laser scanner or multiple cameras, or a lighting dome with more than a hundred light sources. In contrast, this new technique only needs a single camera, to produce high-quality outputs.

“Many traditional methods using a single camera can capture only the 3D geometry of objects, but not the complex reflectance of real-world objects, given by the SVBRDF,” notes Kim. SVBRDF, which stands for spatially-varying bidirectional reflectance distribution functions, is key in obtaining an object’s real-world shape and appearance. “Using only 3D geometry cannot reproduce the realistic appearance of the object in the AR/VR environment. Our technique can capture high-quality 3D geometry as well as its material appearance so that the objects can be realistically rendered in any virtual environment.”

The group demonstrated their framework using a digital camera, the Nikon D7000 and the built-in camera of an Android mobile phone, in a series of examples in their paper, “Practical SVBRDF Acquisition of 3D Objects with Unstructured Flash Photography.” The novel algorithm, which does not require any input geometry of the target object, successfully captured the geometry and appearance of 3D objects with basic, flash photography and reproduced consistent results. Examples that were showcased in the work included diverse set of objects that spanned a wide range of geometries and materials, including metal, wood, plastic, ceramic, resin and paper, and comprised of complex shapes like a finely detailed mini-statute of Nefertiti.

In future work, the researchers hope to further simplify the capture process or extending the method to include dynamic geometry or larger scenes, for instance.

Story Source:

Materials provided by Association for Computing Machinery. Note: Content may be edited for style and length.

Photonic sensors convey information with light instead of electric currents in wires. They can measure, transmit and manipulate streams of photons, typically through optical fibers, and are used to gauge pressure, temperature, distance, magnetic fields, environmental conditions and more.

They are attractive because of their small size, low power consumption and tolerance of environmental variables such as mechanical vibration. But the general consensus has been that high levels of radiation would modify the optical properties of their silicon, leading to incorrect readings.

So NIST, long a world leader in many areas of photonics research, launched a program to answer those questions. The test results indicate the sensors could be customized for measuring radiation dose in both industrial applications and clinical radiotherapy. The results of its first round of testing is reported in Nature Scientific Reports.

Specifically, the NIST results suggest the sensors could be used to track levels of ionizing radiation (with energy high enough to alter the structure of atoms) used in food irradiation to destroy microbes and in medical device sterilization — estimated to be a $7 billion annual market in the U.S. alone. The sensors also have potential applications in medical imaging and therapy, which together are projected to total nearly $50 billion in annual value worldwide by 2022.

“When we looked at publications on the subject, different labs were getting dramatically different results,” said project scientist Zeeshan Ahmed, who is part of NIST’s Photonic Dosimetry Project and leader of NIST’s cutting-edge Photonic Thermometry Project. “That was our main motivation for doing our experiment.”

“Another motivation was the growing interest in deploying photonic sensors that can function accurately in very harsh environments, such as close to nuclear reactors, where radiation damage is a major concern,” Ahmed said. “In addition, the space industry needs to know how these devices would function in high-radiation environments,” said project scientist Ronald Tosh. “Are they going to get damaged or not? What this study shows is that for a certain class of devices and radiation, the damage is negligible.”

“We found that oxide-coated silicon photonic devices can withstand radiation exposure up to 1 million gray,” said Photonic Dosimetry project leader Ryan Fitzgerald, using the SI unit for absorbed radiation. One gray represents one joule of energy absorbed by one kilogram of mass, and 1 gray corresponds to 10,000 chest X-rays. This is roughly what a sensor would receive at a nuclear power plant.

“It’s the upper limit of what our calibrations customers care about,” Fitzgerald said. “So the devices can be assumed to work reliably at industrial or medical radiation levels that are hundreds or thousands of times lower.” Food irradiation, for example, ranges from a few hundred to a few thousand gray, and is typically monitored by its effects on pellets of alanine, an amino acid that changes its atomic properties when exposed to ionizing radiation.

To determine the effects of radiation, the NIST researchers exposed two kinds of silicon photonic sensors to hours of gamma radiation from cobalt-60, a radioactive isotope. In both types of sensors, small variations in their physical properties change the wavelength of the light that travels through them. By measuring those changes, the devices can be used as highly sensitive thermometers or strain gauges. This remains true in extreme environments like space flight or nuclear reactors, only if they continue to function properly under exposure to ionizing radiation.

“Our results show that these photonic devices are robust in even extreme radiation environments, which suggests they could be also used to measure radiation via its effects on physical properties of irradiated devices,” Fitzgerald said. “That should come as good news for U.S. manufacturing, which is anxious to serve the large and growing market for precise delivery of radiation at very small length scales. Photonic sensors could then be developed to measure low-energy electron and X-ray beams used in medical device sterilization and food irradiation.”

They will also be of great interest to clinical medicine, in which physicians strive to treat cancers and other conditions with the lowest effective levels of radiation focused on the smallest dimensions to avoid affecting healthy tissue, including electron, proton and ion beams. Reaching that goal demands radiation sensors with extraordinarily high sensitivity and spatial resolution. “Eventually, we hope to develop chip-scale devices for industrial and medical applications that can determine absorbed dose gradients over distances in the range of micrometers and thus provide unprecedented detail in measurements,” said project scientist Nikolai Klimov. A micrometer is a millionth of a meter. A human hair is about 100 micrometers wide.

The team’s results may have large implications for new medical therapies that employ extremely narrow beams of protons or carbon ions and medical sterilization processes that use low-energy beams of electrons. “Our sensors are naturally small and chip-scale,” Fitzgerald said. “Current dosimeters are on the order of millimeters to centimeters, which can give erroneous readings for fields that vary over those dimensions.”

In the next stage of the research, the team will test arrays of sensors simultaneously in identical conditions to see if variations in dose over small distances can be resolved.

The artificial-intelligence breakthrough, detailed in a paper published in the scientific journal Methods in Ecology and Evolution, is described as a significant advancement in the study and conservation of wildlife. The computer model is now available in a software package for Program R, a widely used programming language and free software environment for statistical computing.

“The ability to rapidly identify millions of images from camera traps can fundamentally change the way ecologists design and implement wildlife studies,” says the paper, whose lead authors are recent UW Department of Zoology and Physiology Ph.D. graduate Michael Tabak and Ryan Miller, both of the U.S. Department of Agriculture’s Center for Epidemiology and Animal Health in Fort Collins, Colo.

The study builds on UW research published earlier this year in the Proceedings of the National Academy of Sciences (PNAS) in which a computer model analyzed 3.2 million images captured by camera traps in Africa by a citizen science project called Snapshot Serengeti. The artificial-intelligence technique called deep learning categorized animal images at a 96.6 percent accuracy rate, the same as teams of human volunteers achieved, at a much more rapid pace than did the people.

In the latest study, the researchers trained a deep neural network on Mount Moran, UW’s high-performance computer cluster, to classify wildlife species using 3.37 million camera-trap images of 27 species of animals obtained from five states across the United States. The model then was tested on nearly 375,000 animal images at a rate of about 2,000 images per minute on a laptop computer, achieving 97.6 percent accuracy — likely the highest accuracy to date in using machine learning for wildlife image classification.

The computer model also was tested on an independent subset of 5,900 images of moose, cattle, elk and wild pigs from Canada, producing an accuracy rate of 81.8 percent. And it was 94 percent successful in removing “empty” images (without any animals) from a set of photographs from Tanzania.

The researchers have made their model freely available in a software package in Program R. The package, “Machine Learning for Wildlife Image Classification in R (MLWIC),” allows other users to classify their images containing the 27 species in the dataset, but it also allows users to train their own machine learning models using images from new datasets.

The lead author of the PNAS article, recent UW computer science Ph.D. graduate Mohammad Sadegh (Arash) Norouzzadeh, is one of multiple contributors to the new paper in Methods in Ecology and Evolution. Other participating researchers from UW are Department of Computer Science Associate Professor Jeff Clune and postdoctoral researcher Elizabeth Mandeville of the Wyoming Cooperative Fish and Wildlife Research Unit.

Other organizations represented in the research group are the USDA’s National Wildlife Research Center, Arizona State University, California’s Tejon Ranch Conservancy, the University of Georgia, the University of Florida, Colorado Parks and Wildlife, the University of Saskatchewan and the University of Montana.

Story Source:

Materials provided by University of Wyoming. Note: Content may be edited for style and length.

“Jet fuel and aviation gasoline are easy to store on an airplane. They are compact and lightweight when compared to the amount of energy they provide. Unfortunately, the actual combustion process is very inefficient. We’re harnessing only a small fraction of that energy but we currently don’t have electrical storage systems that can compete with that,” said Phillip Ansell, assistant professor in the Department of Aerospace Engineering in the College of Engineering at the University of Illinois.

Ansell said adding more batteries to fly farther may seem logical, but it works against the goal to make an aircraft as lightweight as possible. “That’s one of the big barriers we run into when designing battery-powered electrified aircraft. The current technology has very significant range disadvantages. But strong fuel-burn advantages.”

He, along with former aerospace undergraduate student, Tyler Dean, and current doctoral student Gabrielle Wroblewski, utilized a series of simulations to model the performance of hybrid-electric aircraft.

“We started with an existing twin-engine aircraft and looked at how we might create a hybrid-electric drivetrain for it using existing off-the-shelf hardware,” Ansell said. “We wanted to know how well it would perform. If I used a certain set of drivetrain components, I want to know how far the aircraft could fly, how much fuel does it burn, how fast can if climb — all of the overall flight performance changes.”

A flight-performance simulator was created to accurately represent the true flight performance of a Tecnam P2006T on a general mission to include take off, climb, cruise, descent, and landing, along with sufficient reserves to meet FAA regulations. Transition segments were incorporated into the simulation during climb and descent where the throttle setting, flap deployment, propeller rotation rate, and all other flight control variables were either set to mimic input from a typical pilot or prescribed in accordance with the aircraft flight manual.

After configuring the simulator to collect baseline performance data, a parallel hybrid drivetrain was integrated into the simulation. The researchers compared the sensitivity of range and fuel economy to the level of electrification, battery specific energy density, and electric motor power density. The same sensitivities were studied with a series hybrid-electric drivetrain.

Ansell said that, overall, a hybrid-electric drivetrain can lead to substantial improvements in fuel efficiency of a given aircraft configuration, though these gains depend strongly on the coupled variations in the degree of drivetrain electrification and the required mission range. Both of these factors influence the weight allocation of battery and fuel systems, as well as the weight scaling imposed by internal combustion engine and electrical motor components. In general, to obtain the greatest fuel efficiency a hybrid architecture should be used with as much electrification in the drivetrain as is permissible within a given range requirement.

The fuel efficiency improvements were shown to particularly shine for short-range missions, which is a good thing since range limitations serve as one of the key bottlenecks in hybrid aircraft feasibility. Though, through this study the changes in the range capabilities of the aircraft were also able to be forecast with advancements in hybrid component technologies. “For example,” Ansell said, “the propulsion system today could be configured to have 25 percent of its propulsive power come from an electric motor. However, it would only be able to fly about 80 nautical miles. Fast forward to projections for lighter battery technologies for roughly the year 2030 and the same aircraft could fly two and a half to three times as far. The range increase is nonlinear, so the largest improvements can be seen for the most immediate improvements with battery specific energy density, with gradually diminishing returns for that same proportional increase in specific energy.”

“One interesting and unexpected result we observed, however, came about when comparing the parallel and series hybrid architectures. Since the parallel architecture mechanically couples the shaft power of the engine and motor together, only one electrical machine is needed. For the series architecture, a generator is also needed to convert the engine power to electrical power, along with a larger motor than the parallel hybrid configuration to drive the propulsor. Unexpectedly, this aspect made the parallel architecture more beneficial for improved range and fuel burn almost across the board due to its lighter weight. However, we did observe that if significant improvements are made in maturing electrical motor components in the very long term, we may actually someday see better efficiency out of series-hybrid architectures, as they permit a greater flexibility in the placement and distribution of propulsors.”

The team chose to model the Tecnam P2006T using a series of performance variables found in published articles by the aircraft manufacturer. They selected that particular aircraft, in part, because NASA has been working on their X-57 aircraft, which has leading-edge propellers for high lift. “This study was being conducted for NASA, and use of this aircraft also allowed our results to be better applicable to the X-57 concept vehicle,” Ansell said. “Using our data, they will be able to have at least a ballpark idea about how the hybrid system will perform without the other distributed propulsion modifications.”

Ansell said propulsion electrification is still very much an unknown in terms of how a vehicle should be built, engineered, flown. “Our study helps inform those discussions. We looked only at battery storage systems though there are many more that can be implemented, each with their own advantages and disadvantages. This study allowed us to look at what types of advancements need to be made in motor technology, in battery technology, etc.”

The study, “Mission Analysis and Component-Level Sensitivity Study of Hybrid-Electric General Aviation Propulsion Systems,” was conducted by Tyler Dean, Gabrielle Wroblewski, and Phillip Ansell. It appears in the Journal of Aircraft.

This project was supported by NASA Neil A. Armstrong Flight Research Center under Small Business Technology Transfer in collaboration with Rolling Hills Research Corporation.