A free software platform uses crowdsourcing and facial recognition software to identify Civil War soldiers

Oliver Croxton, pictured above right, is computer science Assistant Professor Kurt Luther’s great-great-great uncle. Photo courtesy of the Ken Turner Collection.

Kurt Luther, Virginia Tech assistant professor of computer science, has developed a free software platform that uses crowdsourcing to significantly increase the ability of algorithms to identify faces in photos.

Through the software platform, called Photo Sleuth, Luther seeks to uncover the mysteries of the nearly 4 million photographs of Civil War-era images that may exist in the historical record.

Luther will present his research surrounding the Photo Sleuth platform on March 19 at the Association for Computing Machinery’s Intelligent User Interfaces conference in Los Angeles, California. He will also demonstrate Photo Sleuth at the grand opening of the expanded American Civil War Museum, in Richmond, Virginia, on May 4, 2019.

Luther, a history buff himself, was inspired to develop the software for Civil War Photo Sleuth in 2013 while visiting the Heinz History Center’s exhibit called “Pennsylvania’s Civil War” in Pittsburgh, Pennsylvania. There he stumbled upon a Civil War-era portrait of Oliver Croxton, his great-great-great uncle who served in Company E of the 134th Pennsylvania, clad in a corporal’s uniform.

“Seeing my distant relative staring back at me was like traveling through time,” said Luther.  “Historical photos can tell us a lot about not only our own familial history but also inform the historical record of the time more broadly than just reading about the event in a history book.”

The Civil War Photo Sleuth project, funded primarily by the National Science Foundation, was officially launched as a web-based platform at the National Archives in Washington, D.C., on Aug. 1, 2018, and allows users to upload photos, tag them with visual cues, and connect them to profiles of Civil War soldiers with detailed records of military history. Photo Sleuth’s initial reference database contained more than 15,000 identified Civil War soldier portraits from public domain sources like the U.S. Military History Institute and other private collections.

Prior to the project’s official launch in August, the software platform won the $25,000 Microsoft Cloud AI Research Challenge and the Best Demo Award at the Human Computation and Crowdsourcing 2018 conference in Zurich, Switzerland, for Luther and his team, which includes academic and historical collaborators, the Virginia Center for Civil War Studies, and Military Images magazine.

According to Luther, the key to the site’s post-launch success has been the ability to build a strong user community. More than 600 users contributed more than 2,000 Civil War photos to the website in the first month after the launch, and roughly half of those photos were unidentified. Over 100 of these unknown photos were linked to specific soldiers, and an expert analysis found that over 85 percent of these proposed identifications were probably or definitely correct. Presently, the database has grown to over 4,000 registered users and more than 8,000 photos.

“Typically, crowdsourced research such as this is challenging for novices if users don’t have specific knowledge of the subject area,” said Luther. “The step-by-step process of tagging visual clues and applying search filters linked to military service records makes this detective work more accessible, even for those that may not have a deeper knowledge of Civil War military history.”

Person identification tasks can be challenging in larger candidate pools because there is a larger risk for false positives. The novel approach behind Civil War Photo Sleuth is based on the analogy of finding a needle in a haystack. The data pipeline has three haystack-related components: building the haystack, narrowing down the haystack, and finding the needle in the haystack. When combined, they allow users to identify unknown soldiers while reducing the risk of false positives.

Building the haystack is done by incentivizing users to upload scanned images of the fronts and backs of Civil War photos. Any time a user uploads a photo to identify it, the photo gets added to the site’s digital archive or “haystack,” making it available for future searches.

Following upload, the user tags metadata related to the photograph such as photo format or inscriptions, as well as visual clues, such as coat color, chevrons, shoulder straps, collar insignia, and hat insignia. These tags are linked to search filters to prioritize the most likely matches. For example, a soldier tagged with the “hunting horn” hat insignia would suggest potential matches who served in the infantry, while hiding results from the cavalry or artillery. Next, the site uses state-of-the-art face recognition technology to eliminate very different-looking faces and sort the remaining ones by similarity. Both the tagging and face recognition steps narrow down the haystack.

Finally, users find the needle in the haystack by exploring the highest-probability matches in more detail. A comparison tool with pan and zoom controls helps users carefully inspect a possible match and, if they decide it’s a match, link the previously unknown photo to its new identity and biographical details.

The military records used by the filters come from myriad public sources, including the National Park Service Soldiers and Sailors Database.

Retracing historical Civil War photos through facial recognition software like Photo Sleuth has broad applications beyond identifying historical photos, too. The software has the potential to generate new ways to think about building person identification systems that look beyond face recognition and leverage the complementary strengths of both human and artificial intelligence.

Learn more: Researcher releases facial recognition software to identify Civil War soldiers

 

 

The Latest on: Software platform that uses crowdsourcing

via  Bing News

 

A process to 3D print piezoelectric materials

Internal topology of 3D printed piezoelectrics spanning the width of human hair

The piezoelectric materials that inhabit everything from our cell phones to musical greeting cards may be getting an upgrade thanks to work discussed in the journal Nature Materials released online Jan 21.

Xiaoyu ‘Rayne’ Zheng, assistant professor of mechanical engineering in the College of Engineering, and a member of the Macromolecules Innovation Institute, and his team have developed methods to 3D print piezoelectric materials that can be custom-designed to convert movement, impact and stress from any directions to electrical energy.

“Piezoelectric materials convert strain and stress into electric charges,” Zheng explained.

The piezoelectric materials come in only a few defined shapes and are made of brittle crystal and ceramic – the kind that require a clean room to manufacture. Zheng’s team has developed a technique to 3D print these materials so they are not restricted by shape or size. The material can also be activated – providing the next generation of intelligent infrastructures and smart materials for tactile sensing, impact and vibration monitoring, energy harvesting, and other applications.

Unleash the freedom to design piezoelectrics

Piezoelectric materials were originally discovered in the 19th century. Since then the advances in manufacturing technology has led to the requirement of clean-rooms and a complex procedure that produces films and blocks which are connected to electronics after machining. The expensive process and the inherent brittleness of the material, has limited the ability to maximize the material’s potential.

Zheng’s team developed a model that allows them to manipulate and design arbitrary piezoelectric constants, resulting in the material generating electric charge movement in response to incoming forces and vibrations from any direction, via a set of 3D printable topologies. Unlike conventional piezoelectrics where electric charge movements are prescribed by the intrinsic crystals, the new method allows users to prescribe and program voltage responses to be magnified, reversed or suppressed in any direction.

“We have developed a design method and printing platform to freely design the sensitivity and operational modes of piezoelectric materials,” Zheng said. “By programming the 3D active topology, you can achieve pretty much any combination of piezoelectric coefficients within a material, and use them as transducers and sensors that are not only flexible and strong, but also respond to pressure, vibrations and impacts via electric signals that tell the location, magnitude and direction of the impacts within any location of these materials.”

3D printing of piezoelectrics, sensors and transducers

A factor in current piezoelectric fabrication is the natural crystal used. At the atomic level, the orientation of atoms are fixed. Zheng’s team has produced a substitute that mimics the crystal but allows for the lattice orientation to be altered by design.

“We have synthesized a class of highly sensitive piezoelectric inks that can be sculpted into complex three-dimensional features with ultraviolet light. The inks contain highly concentrated piezoelectric nanocrystals bonded with UV-sensitive gels, which form a solution – a milky mixture like melted crystal – that we print with a high-resolution digital light 3D printer,” Zheng said.

The team demonstrated the 3D printed materials at a scale measuring fractions of the diameter of a human hair. “We can tailor the architecture to make them more flexible and use them, for instance, as energy harvesting devices, wrapping them around any arbitrary curvature,” Zheng said. “We can make them thick, and light, stiff or energy-absorbing.”

The material has sensitivities 5-fold higher than flexible piezoelectric polymers. The stiffness and shape of the material can be tuned and produced as a thin sheet resembling a strip of gauze, or as a stiff block. “We have a team making them into wearable devices, like rings, insoles, and fitting them into a boxing glove where we will be able to record impact forces and monitor the health of the user,” said Zheng.

“The ability to achieve the desired mechanical, electrical and thermal properties will significantly reduce the time and effort needed to develop practical materials,” said Shashank Priya, associate VP for research at Penn State and former professor of mechanical engineering at Virginia Tech.

New applications

The team has printed and demonstrated smart materials wrapped around curved surfaces, worn on hands and fingers to convert motion, and harvest the mechanical energy, but the applications go well beyond wearables and consumer electronics. Zheng sees the technology as a leap into robotics, energy harvesting, tactile sensing and intelligent infrastructure, where a structure is made entirely with piezoelectric material, sensing impacts, vibrations and motions, and allowing for those to be monitored and located. The team has printed a small smart bridge to demonstrate its applicability to sensing the locations of dropping impacts, as well as its magnitude, while robust enough to absorb the impact energy. The team also demonstrated their application of a smart transducer that converts underwater vibration signals to electric voltages.

“Traditionally, if you wanted to monitor the internal strength of a structure, you would need to have a lot of individual sensors placed all over the structure, each with a number of leads and connectors,” said Huachen Cui, a doctoral student with Zheng and first author of the Nature Materials paper. “Here, the structure itself is the sensor – it can monitor itself.”

Learn more: Mechanical engineers develop process to 3D print piezoelectric materials

 

 

The Latest on: 3D printed piezoelectrics

via  Bing News

 

 

 

A new bacteria-based drug delivery system

Description of graphical element: NanoBEADS agents are constructed by conjugating poly(lactic?co?glycolic acid) nanoparticles with tumor?targeting Salmonella typhimurium. NanoBEADS enhance retention and distribution of nanoparticles in solid tumors by up to a remarkable ?100?fold, through intercellular (between cells) self?replication and translocation. This transport enhancement is achieved autonomously, without the need for any externally applied driving force or control input.

An interdisciplinary team of three Virginia Tech faculty members affiliated with the Macromolecules Innovation Institute has created a drug delivery system that could radically expand cancer treatment options.

The conventional cancer treatment method of injecting nanoparticle drugs into the bloodstream results in low efficacy. Due to the complexities of the human body, very few of those nanoparticles actually reach the cancer site, and once there, there’s limited delivery across the cancer tissue.

The new system created at Virginia Tech is known as Nanoscale Bacteria-Enabled Autonomous Drug Delivery System (NanoBEADS). Researchers have developed a process to chemically attach nanoparticles of anti-cancer drugs onto attenuated bacteria cells, which they have shown to be more effective than the passive delivery of injections at reaching cancer sites.

NanoBEADS has produced results in both in vitro (in tumor spheroids) and in vivo (in living mice) models showing up to 100-fold improvements in the distribution and retention of nanoparticles in cancerous tissues.

This is a product of the five-year National Science Foundation CAREER Award of Bahareh Behkam, associate professor of mechanical engineering. Collaborators on this interdisciplinary team are Rick Davis, professor of chemical engineering, and Coy Allen, assistant professor of biomedical sciences and pathobiology in the Virginia-Maryland College of Veterinary Medicine.

“You can make the most amazing drugs, but if you cannot deliver it where it needs to go, it cannot be very effective,” Behkam said. “By improving the delivery, you can enhance efficacy.”

This work, which combines expertise in mechanical engineering, biomedical engineering, chemical engineering, and veterinary medicine, was recently detailed in Advanced Science.

Using salmonella for good

Humans have noticed, even as far back as Ancient Egypt, that cancer went into remission if the patient also contracted an infection like salmonella. Neither are ideal, but humans can treat salmonella infections more effectively than cancer.

In modern times, Allen said the idea of treating cancer with infections traces back to the late 1800s and has evolved into immunotherapy, in which doctors try to activate the immune system to attack cancerous cells.

Of course, salmonella is harmful to humans, but a weakened version could in theory provide the benefits of immunotherapy without the harmful effects of salmonella infection. The concept is similar to humans receiving a weakened flu virus in a vaccine to build immunity.

Over six years ago, Behkam came up with the idea of augmenting bacterial immunotherapy to also attack cancer with conventional anti-cancer drugs. The problem was the passive delivery of anti-cancer drugs doesn’t work very well.

Given her background in bio-hybrid microrobotics, she wanted to use salmonella bacteria as autonomous vehicles to transport the medicine, in nanoparticle form, directly to the cancer site.

The work began with Behkam’s first doctoral student, Mahama Aziz Traore, constructing the first generation of NanoBEADS by assembling tens of polystyrene nanoparticles onto E. colibacteria. After thoroughly studying the dynamics and control aspects of the NanoBEADS systems for a few years, Behkam brought Davis into the project because he had experience creating polymer nanoparticles for drug delivery.

“She mentioned this radically different approach for delivering drugs and nanoparticles,” Davis said. “I walked away from the conversation thinking, ‘Man, if this thing could work, it would be fantastic.’”

Behkam chose this particular bacterial strain, Salmonella enterica serovar Typhimurium VNP20009, because it has been thoroughly studied and successfully tested in a phase one clinical trial.

“Its (salmonella’s) job as a pathogen is to penetrate through the tissue,” Behkam said. “What we thought is if bacteria are so good at moving through the tissue, how about coupling nanomedicine with the bacterium to carry that medicine much farther than it’d passively diffuse on its own?”

Trial and error

Although Behkam had a vision for the new drug delivery system, it took several years for it to become reality.

“The process of creating nanoparticles and then attaching them to bacteria in a robust and repeatable manner was challenging, but add on top of that ensuring the bacteria stay alive, discovering the mechanism of bacteria transport in cancerous tissue, and devising ways to quantitatively describe the effectiveness of NanoBEADS, and this was a difficult project,” Davis said.

SeungBeum Suh, Behkam’s former Ph.D. student, and Amy Jo, Davis’ former Ph.D. student, worked together on attaching nanoparticles while keeping the bacteria alive. It wasn’t until their fourth attempt that they started finding success.

“We collaborated to make these particles, and we attached them to the bacteria,” Behkam said. “Then the question was what is the mechanism of their translocation in the tumor? How far do they go into the tumor? How do we present a quantitative measure of their performance?”

Behkam along with Suh and current doctoral student Ying Zhan tested their nanoparticle-attached salmonella in lab-grown tumors. They found up to 80-fold improvements in nanoparticle penetration and distribution using the NanoBEADS platform, compared to passively diffusing nanoparticles.

Furthermore, Suh and Behkam found out that NanoBEADS largely penetrate the tumor by translocating through the space in between cancer cells.

Behkam wanted to strengthen the NanoBEADS results past the in vitro stage. With a top-flight veterinary school down the road, she enlisted Allen, her fellow MII faculty member, to test the NanoBEADS system in vivo. Tests in breast cancer tumors in mice produced results showing significant improvements compared to passive delivery.

The tests showed that there was about 1,000 times more salmonella cells in the tumor compared to the liver and 10,000 times more than the spleen.

“Most notably, the salmonella itself helped keep the particles in the tumor up to 100-fold better, which would suggest it would be an effective delivery vehicle,” Allen said.

The next step in the research is to load cancer therapeutics into the NanoBEADS system to test the potential enhancement in efficacy.

From bench to kennel to bedside

The collaboration highlights the diversity of interdisciplinary research possible through MII and Virginia Tech.

“The synergistic integration of diverse expertise has been essential to the high-impact discoveries that resulted from this work,” Behkam said.

With the addition of the Virginia Tech Carilion School of Medicine and Fralin Biomedical Research Institute at VTC, Allen said Virginia Tech has the possibility to test scientific research “from bench to kennel to bedside.”

“The project could not move forward without each of the three parts,” Allen said. “The study would not have gotten into such a high impact journal without having the chemistry, the background of the pathogen, the idea, and having the physiological and clinical relevance of testing it in an actual tumor in an actual animal model.”

Davis said all drug delivery mechanisms have to go through animal trials, so having an “absolutely fantastic” college of veterinary medicine on campus took the research to a higher level.

“One thing that attracted me to this project was the ability to work with people like Bahareh and Coy who work with cells and animal studies to really translate the work,” Davis said. “It’s hard to find that combination of people in a lot of schools.”

Learn more: Virginia Tech researchers create a bacteria-based drug delivery system that outperforms conventional methods

 

 

The Latest on: Bacteria-based drug delivery system

via  Bing News

 

Getting much closer to plastics that can be produced to replace today’s petroleum plastics that persist in landfills and oceans

The setup of the photoredox polymerization in the glove box with a cold trap, irradiated by the blue LED light.

There’s a good chance you’ve touched something made out of the polyolefin polymer today. It’s often used in polyethylene products like plastic bags or polypropylene products like diapers.

As useful as polyolefins are in society, they continue to multiply as trash in the environment. Scientists estimate plastic bags, for example, will take centuries to degrade.

But now, researchers at Virginia Tech have synthesized a biodegradable alternative to polyolefins using a new catalyst and the polyester polymer, and this breakthrough could eventually have a profound impact on sustainability efforts.

Rong Tong, assistant professor in the Department of Chemical Engineering and affiliated faculty member of Macromolecules Innovation Institute (MII), led the team of researchers, whose findings were recently published in the journal Nature Communications.

One of the largest challenges in polymer chemistry is controlling the tacticity or the stereochemistry of the polymer. When multiplying monomer subunits into the macromolecular chain, it’s difficult for scientists to replicate a consistent arrangement of side-chain functional groups stemming off the main polymer chain. These side-chain functional groups greatly affect a polymer’s physical and chemical properties, such as melting temperature or glass-transition temperature, and regular stereochemistry leads to better properties.

Tong said his group has now found a way to create regular stereochemistry with polyesters.

“There’s no method available to do this kind of chemistry,” Tong said. “People have done similar work with polylactide before, but we’ve fundamentally shown that if we control the stereochemistry, the polyesters will have improved physical and chemical properties.”

Tong and his postdoc, Quanyou Feng, combined a new photoredox Ni/Ir catalyst — a surprisingly simple chemical process that uses a household light bulb to start the reaction — with a stereoselective Zn catalyst to initiate the ring-opening polymerization of the O-carboxyanhydride monomer to create these improved polyesters. The monomers can be conveniently polymerized within just a few hours with trace amounts of catalysts. The resulting material has a high molecular weight, thermal stability and crystallinity, and can degrade in basic water solution.

“If you use a regular catalyst, it doesn’t have stereochemistry control, but we found that our catalyst can do that,” Tong said. “In our paper, we demonstrate how to design such stereoselective catalysts and how they help with stereochemistry control.”

O-carboxyanhydrides are made out of amino acids, which are natural organic compounds, so these polyesters would degrade, unlike the current nondegradable polyolefins. In addition, O-carboxyanhydrides can bring different functional groups to the polyester and diversify the polymer’s application. Currently, the FDA has only approved a few polyesters for biomedical application.

After finalizing the synthesis, Tong then worked with Guoliang “Greg” Liu, an assistant professor in the Department of Chemistry and fellow affiliated faculty member with MII, to show that the new polymers had improved properties.

“Dr. Tong’s lab has outstanding catalyst design and polymerization techniques, and we have excellent characterization and processing skillsets, so it’s natural for us to work together,” Liu said. “Controlling and proving tacticity is not a trivial process. Using differential scanning calorimetry and nuclear magnetic resonance, we provide strong evidence for the structure and properties that we’re going for.”

Developing these polyesters into applications is still down the line, but Liu said for now this is a significant advancement for materials research.

“This polyester synthesis that controls the tacticity can provide a new library of polymer materials that we haven’t had before,” Liu said.

This piece of innovative chemistry has Tong and Liu excited for a future that degradable and green plastics can be produced to replace today’s petroleum plastics that persist in landfills and oceans for decades or centuries.

Tong mentioned that this new polymer synthesis technology has been demonstrated only at the academic lab scale. There is still much work to be done to characterize these functional materials and perfect the patent-pending synthesis scale-up process.

“It would be our dream to see these degradable polyesters materialize in the marketplace, for both the plastic industry and biomedical application,” Tong said.

Learn more: Polymer researchers discover path to sustainable and biodegradable polyesters

 

 

The Latest on: Biodegradable plastics

via  Bing News

 

Fog “harps” could collect more than three-times the amount of clean water from fog

To test the fog harp’s design, researchers constructed small-scale models of vertical wire arrays that could be placed inside an environmental chamber with artificial fog. The team discovered that water collection efficiency continued to increase with smaller and smaller wires.

Fog harvesting may look like whimsical work.

After all, installing giant nets along hillsides and mountaintops to catch water out of thin air sounds more like folly than science. However, the practice has become an important avenue to clean water for many who live in arid and semi-arid climates around the world.

A passive, durable, and effective method of water collection, fog harvesting consists of catching the microscopic droplets of water suspended in the wind that make up fog. Fog harvesting is possible – and has gained traction over the last several decades – in areas of Africa, South America, Asia, the Middle East, and even California. As illustrated by recent headlines of South Africa’s countdown to “Day Zero,” or the day the water taps are expected to run dry, water scarcity continues to be a growing problem across the globe. Leading researchers now estimate that two-thirds of the world’s population already live under conditions of severe water scarcity at least one month of the year.

Fog harvesting could help alleviate that shortage, and now an interdisciplinary research team at Virginia Tech has improved the traditional design of fog nets to increase their collection capacity by threefold.

Published in ACS Applied Materials & Interfaces and partially funded by the Virginia Tech Institute for Creativity, Arts, and Technology, the team’s research demonstrates how a vertical array of parallel wires may change the forecast for fog harvesters. In a design the researchers have dubbed the “fog harp,” these vertical wires shed tiny water droplets faster and more efficiently than the traditional mesh netting used in fog nets today.

“From a design point of view, I’ve always found it somewhat magical that you can essentially use something that looks like screen door mesh to translate fog into drinking water,” said Brook Kennedy, associate professor of industrial design in the College of Architecture and Urban Studies and one of the study’s co-authors. “But these parallel wire arrays are really the fog harp’s special ingredient.”

Fog nets have been in use since the 1980s and can yield clean water in any area that experiences frequent, moving fog. As wind moves the fog’s microscopic water droplets through the nets, some get caught on the net’s suspended wires. These droplets gather and merge until they have enough weight to travel down the nets and settle into collection troughs below. In some of the largest fog harvesting projects, these nets collect an average of 6,000 liters of water each day.

However, the traditional mesh design of fog nets has long posed a dual constraint problem for scientists and engineers. If the holes in the mesh are too large, water droplets pass through without catching on the net’s wires. If the mesh is too fine, the nets catch more water, but the water droplets clog up the mesh without running down into the trough and wind no longer moves through the nets.

Thus, fog nets aim for a middle ground, a Goldilocks zone of fog harvesting: mesh that’s not too big and not too small. This compromise means nets can avoid clogging, but they’re not catching as much water as they could be.

“It’s an efficiency problem and the motivation for our research,” said Jonathan Boreyko, assistant professor in the Department of Biomedical Engineering and Mechanics in the College of Engineering. As a co-author of the study, Boreyko consulted on the theory and physical aspects of the fog harp’s design.

“That hidden regime of making the wires smaller but not clogging is what we were trying to accomplish. It would be the best of both worlds,” he said.

Since the water droplets caught in a fog net move downward with gravity, Boreyko hypothesized that removing the horizontal wires of the net would alleviate some of the clogging. Meanwhile, Kennedy, who specializes in biomimetic design, found his inspiration for the fog harp in nature.

“On average, coastal redwoods rely on fog drip for about one-third of their water intake,” said Kennedy. “These sequoia trees that live along the California coast have evolved over long periods of time to take advantage of that foggy climate. Their needles, like those of a traditional pine tree, are organized in a type of linear array. You don’t see cross meshes.”

Mark Anderson, a study co-author and then-undergraduate student in the Department of Mechanical Engineering, built several scale models of the fog harp with varying sizes of wires. Weiwei Shi, a doctoral student in the engineering mechanics doctoral program as well as the study’s lead author, tested the small prototypes in an environmental chamber and developed a theoretical model of the experiment.

“We found that the smaller the wires, the more efficient the water collection was,” said Boreyko. “These vertical arrays kept catching more and more fog, but the clogging never happened.”

The team has already constructed a larger prototype of the fog harp – a vertical array of 700 wires that measures 3 feet by 3 feet – in an effort led by Kennedy with assistance from Josh Tulkoff, study co-author and a then-undergraduate student in the industrial design program. They plan to test the prototype on nearby Kentland Farm.

Through its unique combination of science and design, the researchers hope the fog harp will one day make a big impact where it’s needed most – in the bottom of the water bucket.

Learn more: Harvesting water from fog with harps

 

The Latest on: Water harvesting

via  Bing News