Building a Better Workplace with the Whole Body Scanner

When construction workers lift the end frame of a scaffold into place, how much stress do they put on their lower backs? If a safety harness catches a falling worker, will the harness cause as much harm as the fall? Do agriculture workers risk too much injury in common tasks, such as picking apples or operating a tractor?

The answers to these questions depend on the size and shape of America's construction and agriculture workers. When you know a woman's shape, you can evaluate whether her safety harness poses its own safety threat. And when you know the length of a farm worker's reach, you can see how he interacts with apple trees and tractors.

To get the crucial measurements, the National Institute for Occupational Safety and Health (NIOSH) is acquiring a fleet of Cyberware color 3D scanners—including a WB4 Whole Body scanner, a 3030HRC/PS for scanning heads, and a hand scanner. The Cyberware scanners will be installed in a series of new NIOSH labs that will investigate a wide range of safety issues.

According to Dr. Hongwei Hsiao, Acting Chief of the Division of Safety Research's Safety Controls Section, existing data on the size and shape of construction and agriculture workers is thin, at best. Safety researchers have been relying on data gathered from a set of NASA anthropometry source books published in 1978. However, the books mainly collected manual measurements of Air Force personnel.

Researchers have a pretty good idea that construction and agriculture workers are bigger and more muscular than the average airman.

How much bigger are they? The Whole Body scanner offers the ideal way to find out because it captures the shape of the entire human body in one quick pass. Dr. Hsiao plans to scan 5,000 construction works and an equal number of agriculture workers over four years. After NIOSH researchers develop their data-collection and analysis procedures, he intends to install the WB scanner in a truck and take it to work sites. Because of the scanner's speed, subjects will miss only a few minutes of work.

By making a 3D scan of workers, NIOSH will get complete data for biometrics studies. The scan data reveals the size as well as the volume/mass of limbs, so that researchers can calculate the different compression forces that would result if a specific worker performed a specific task in various postures. This kind of analysis can be automated using custom software. The software will analyze the thousands of worker scans in a consistent way so that researchers obtain accurate statistics.

With this information, NIOSH can provide a variety of useful results. For example, the agency can tell protective-gear manufacturers the number of harness sizes needed to accommodate the worker population. If certain tasks put unacceptable stress on workers' spines, NIOSH could recommend that equipment be developed to help.

Such results are critical to on-the-job safety and productivity. In 1990, the Bureau of Labor Statistics estimated that 6.4 million non-fatal traumatic injuries occurred in US workplaces. These injuries resulted in approximately 60 million lost workdays.

NIOSH believes that many of these injuries and lost days can be prevented through a better understanding of how workers interact with their workplaces, tools, and protective equipment. In addition to gathering anthropometric data on the traditional burly construction work, Dr. Hsiao points out that NIOSH needs to assess other workers, such as woman and members of ethnic groups who tend to be smaller than the traditional norm. These workers are increasingly moving into construction and agriculture jobs.

For the first time, the Cyberware Whole Body scanner makes it possible to find out exactly what shape American workers are in. Will the NIOSH scans reinforce Herculean stereotypes? Stay tuned.

The Fastest Zipper Tool in the West

In the previous issue of the newsletter, we previewed the new zippering software created by Phil Dench at the Curtin University of Technology, Perth, Western Australia. This software tool now has a name—CyPie—and it is available from Cyberware in an advanced beta version.

CyPie zippers together multiple scans of an object at breath-taking speeds. It automatically assembles any number of linear scans into a single model and retains any color texture mapping information along with the range data.

To illustrate how CyPie works, we scanned the ceramic doll shown here, using a 3030RGB/HIREZ/MM scanner and a technique that works well with CyPie. Specifically, we created a batch file that directed the MM to take a linear scan of the doll, rotate the doll 10 degrees, then take another linear scan, rotate the doll another 10 degrees, and so on until we had a complete set of 36 scans. Making this series of linear scans takes longer than making one cylindrical scan, but the quality of the linear method is much better—if you use CyPie to assemble the linear scans.

After collecting the 36 scans, we could have cleaned up stray errors in CyPie before zippering the scans together, but in this case it was easier t assemble them first. CyPie took about 10 minutes to do the job, entirely automatically. When CyPie combines two scans, it keeps the polygons from each scan that are most likely to be accurate. On the doll, for example, the middle of the head in each scan is likely to be highly accurate, while the edges are less distinct. CyPie recognizes this difference in confidence levels and chooses polygons accordingly. When the zippering process is complete, the final model consists of the best polygons from each of the original scans.

Although CyPie doesn't contain any polygon editing tools, you can read CyPie files into CyZip for editing. You can also use Cyberware-supplied converters to translate the CyPie file to formats that can be edited in a variety of software environments. The CyPie package also includes PlyView, a viewer that displays the range data, the color map applied to the range data, the CyPie confidence levels, polygon representations, or polylines.

To get a good look at the results described here, you can download PlyView and the doll model from Cyberware's World Wide Web or FTP sites: http://www.cyberware.com, then follow the "CyPie Sample File" link to download. ftp://ftp.cyberware.com/pub/CyPie_Demo_Files.tar.Z.

Bear in mind that CyPie does not have the ability to zipper together any two arbitrary scans of an object. You have to use linear scans, and they have to overlap in known rotational increments. That can make it difficult to capture the top and bottom of an object such as the doll, but there are two ways around this limitation. You can selectively scan such areas and merge them with the rest of the model using CyZip. Or, for somewhat better results, you can place the object on its side, take additional scans of it that way, assemble these scans with CyPie, then merge the resulting model with the other almost-complete model in CyZip. Another new program (CyEat) promises to automate the merging of CyPie models; check the Cyberware Web page for more information. http://www.cyberware.com

Scanning Tips & Tricks

If you see spikes in a scanned image that you cannot attribute to strong light sources, the mirrors on your scanner might be dirty. Fortunately, the mirrors are easy to check and easy to clean—IF YOU FOLLOW THE DIRECTIONS IN THIS ARTICLE. It is important to be careful because even a small scratch on a mirror can impair the operation of your scanner, and mirrors are expensive to replace.

You can quickly check your mirrors for dirt by turning on the scanner's laser using the "view" mode in Echo, then holding a piece of paper so that you can see the laser line on the paper. The line should be thin and crisp, with no extraneous light to either side.

If your mirrors need cleaning, wet a cotton ball with either isopropyl alcohol or methanol. Lightly swab the dirty area. Do not apply any direct pressure on the mirror with your fingers, and take care to avoid rubbing against any of the other mirrors while you are cleaning. NEVER use anything other than a cotton ball, a very soft tissue, or a cotton swab on the mirrors, and never apply hard pressure.

As a final step, check the mirrors again with the flashlight or laser to ensure that the area is now free of dust and smudges.

Scans Reveal the Progress of Healing

Groundbreaking research involving a Cyberware scanner promises to give doctors their first accurate method for assessing the healing of wounds. If that method tests out, it will in turn provide a way to evaluate new drugs and other therapies. As it is, medical researchers have had to rely largely on subjective judgements.

The wound study, led by Barbara McQuiston of Dayton, OH-based Sytronics, Inc., is being funded by the Air Force Surgeon General. McQuiston is working with several groups a Wright-Patterson Air Force Base, including the Computerized Anthropometry Research & Design Laboratory and the Hyperbaric Medicine department.

Doctors find it hard to tell whether a serious wound is healing well, because human wounds can actually expand as they heal. Thus, simply measuring the size of a wound from day to day does not give an accurate gauge of the healing process. And because would healing is so difficult to evaluate, doctors cannot be entirely sure whether a new healing therapy is helping or not.

One exception is hyperbaric therapy. The pressurized oxygen in a hyperbaric chamber helps kill certain bacteria and promotes healing by getting more oxygen to the wound tissues. McQuiston is including hyperbaric therapy in her study, and she speculates that her methods could prove useful for some of this therapy's most high-profile users: pro athletes. The ability to accurately assess a wound's status could make it possible to know exactly when an NFL linebacker is fit to get back in the game.

Attempts to map healing progress using conventional 2D photographs or caliper measurements have not yielded good results, because would are irregular, three-dimensional forms. Researchers have measured the size of would by filling them with a gelatinous material, allowing the material to harden, then dropping the material into water to measure the displacement. All of these methods are crude, tedious, and inaccurate.

McQuiston's original goal for the Cyberware scanner was to get an accurate representation of each wound's borders. Using a 3030/RGB/LN scanner, she captures 3D color models of diabetics' wounds on a regular basis. So far, she has found that the models' color give an excellent signature for a wounds' state. As a wound heals, specific types of tissues gradually grow and give way to other types of tissues. The color texture map clearly shows the progress of these changes in three dimensions—a crucial advantage. Surpassing McQuiston's expectations, the scans reveal the characteristics of the tissues, as well as the wound's borders.

The 3D models form the Cyberware scanner provide such a clear healing signature that McQuiston has been able to map the progress of different healing tissues automatically using the Queen Victoria algorithm from computer vision research. (The algorithm was developed to help match up an old motion picture of Queen Victoria with a separate audio recording of her speech.) This automated method provides medical researchers with their first full objective technique for evaluating wounds.

Along with the Cyberware scanner, McQuiston uses magnetic resonance imaging (MRI) and spectroscopy (MRS, an imaging technique that shows the distribution of any give type of atoms in the body) to evaluate the inner workings of wounds. She aligns each of these scans by making a thermoplastic mold to hold a patient's wounded limb in exactly the same orientation for each scan. She can then overlay the models from the scans using KB Vision or Analyze from the Mayo Clinic. By combining the Cyberware model with the MRI/MRS models, McQuiston can easily see the relationship between a wound's inner and outer features.

McQuiston hopes to collect enough data on healed wounds to feed into a neural net (a circuit that automatically detects the patterns that are common across many examples). With enough data to generalize the healing pattern, the neural net may provide descriptions of how fast a specific wound should heal. This information would prove invaluable in pursuing new therapies.

Cyberware Receives First Place in CGW's Editor's Choice Awards

Cyberware's WB4 and WB2 were recently awarded First Place in Computer Graphics World's 1995 Editor's Choice Awards. The award program is designed to track and review new computer graphics products introduced each year. The competition for the awards was especially intense in 1995 given the high number of products introduced. The award, one of only 75 given nationwide, is an acknowledgment of Cyberware's ongoing commitment to creativity, innovation and quality control.

Lucy Arrives in the Cyber Age

One of the oldest human skulls ever found—dubbed Lucy by paleontologists-has entered the cutting edge of the present via the Cyberware 3030/RGB/PS scanner. Cyberware distributor Cyber_Site Europe recently scanned Lucy, along with 500,000- and 30,000-year-old skulls, for a special exhibit at London's Museum of natural History. The interactive exhibit is among the largest and most ambitious electronic arts projects ever staged in the UK.

Museum visitors see the skulls rendered in black faceted crystal, morphing into one another in evolutionary sequence to show what humans have been up to for the past three million years. The sequence was created by Christian Hogue, founder of the London-based production company Lost in Space, which commissioned Cyber_Site to make the scans.

Return to the Newsletter Listing