The company has incorporated the mother of all photo booths into its stand at the Cebit trade show in Hanover, Germany, this week, and is inviting visitors to step inside to create their own digital avatar.
HPE integrated offerings from three companies to create the Avatar Platform, which builds 3D models of people that can then be digitally clothed and animated.
The avatar creation process begins in a full body scanner that is about 3 meters on a side. Built by Doob Group, it takes 64 pictures simultaneously using an array of cameras stacked in several rings around the subject.
Those images are then passed to the Quantum Human software package from Quantum Matrix, which pieces them together to create a three-dimensional mesh or model of the person in the scanner.
Finally, software from South Korean company Physan is used to simulate virtual clothing. Working from the original manufacturing patterns for the clothing, it can simulate the appearance of the finished article and even its drape -- the way it folds and curves as it hangs from the body.
Although the photos are taken in less than a second, the whole process from entering the room to generating the avatar takes 8 to 10 minutes, said Manuel Meyer of HPE enterprise services.
For now, HPE's system is just a prototype costing about €120,000 (US$135,000).
"We can bring that down to €10,000 to €15,000 by scaling," he said.
Doob already has its scanners in retail stores in New York, Los Angeles, Berlin and its home town, Düsseldorf, where it sells people 3D-printed models of themselves for between €95 and €575 depending on scale. It charges an additional €35 for the data in a 3D-PDF file.
HPE hopes the data from its integrated system can be free for the end user, with other parties subsidizing the cost of the scan.
The company sees three target markets for the system.
First is online clothing retail, where the value of the system would be in reducing the cost of returns, Meyer said.
The average return rate is around 30 percent, according to HPE, which estimates that allowing online customers to "try on" clothes for color and size could reduce that to 10 percent, saving businesses globally up to $36 billion each year.
Using a digital avatar with the same proportions as their own body, customers can test how purchases will look and, to some extent, how they will fit, increasing the chances that they will be satisfied with a purchase.
"For fashion, you need to wear tight clothes for the scan," he said. That requirement could change as HPE is looking into ways to allow people to wear normal clothes during the scan, then take a few key measurements that can be used to adapt a standardized statistical model of the human body to create the model.
Fitness centers could use the system to help customers track the effectiveness of their workouts. "When you start a training program, you have a certain body shape," Meyer said. By taking regular scans, "you can see how it changes over time."
Finally, there's the PC gaming industry, which might pay for the scans as a way to maintain player engagement.
Many online role-playing and first-person shooter games allow players to customize how they see themselves or how others see them. However, the customization is typically limited to a menu of clothing or accessories dictated by the game developer, or the application of a "skin" or pattern to a standard body shape generated by the software.
Enabling digital avatars that mimic their body shape and appearance could be a way to increase player interest.
HPE is talking to game developers about allowing players to upload avatars in file formats commonly used for the exchange of 3D-modelling data, including FBX and OBJ.
"They will be able to use their avatars in games as soon as game manufacturers open up the API," Meyer said.
Once people have their digital avatar, they can reuse the data in other applications, he said.
Further out, the system could be used to create avatars for virtual reality conferencing systems, he said. That could potentially reduce the bandwidth required to convey body movements and facial expressions viewable from any angle.