When most people think of facial image scanning, there’s a good chance their thought process immediately goes to facial recognition technology. After all, that’s the most prevalent use case we hear about today – millions of people are unlocking their smartphones with facial recognition, and anyone who watches crime dramas on TV has seen the technology used many times.
The truth is, while facial recognition is a use case for image scanning, it’s hardly the only one. Countless new use cases are emerging as scanning technology improves – which it must do for more advanced applications that require highly accurate, detailed imaging. Let’s face it, how accurate is the imaging software on your phone if it can be unlocked with a photo, or if a son’s face can unlock his father’s phone?
Therein lies that challenge. How do you take scanning to the next level to be able to create advanced applications for healthcare and other verticals? The truth is, it’s not about the image itself as it is about what you do with the image data.
IKIN has been revolutionizing imaging for many years now with its holography technology. It has brought both small- and large-format headset-free holography to market with a number of use cases. Last year, it demonstrated its technology and officially launched its IKIN University program for content developers at ITEXPO.
IKIN’s holography depends on image scanning data, of course, so when the company was approached to solve a specific problem during the COVID-19 pandemic, it accepted and rose to the challenge.
Cosmetics companies naturally have to put their products through rigorous testing to understand their effects over time. It takes time, and it requires detailed views of test subjects’ faces and skin, a process that became impossible during the pandemic. Sure, high-res images could be taken, but 2D images aren’t able to identify small changes in the soft-body structure of the face – the level of detail required for cosmetics testing, for instance.
One cosmetics firm called on IKIN to help solve the challenge with a solution that would work under any circumstances. IKIN had developed its Cubara dimensional scanning platform that could extract highly accurate image data from high-resolution photos or video. That data is fed into a Convolutional Neural Network (CNN) that breaks down single images to create a 3D point cloud of the face – or any object.
The CNN is utilized to breakdown single images based on a wide pool of pre-trained data for object recognition. Within this system, a customized CNN is generated at the point of initial scan. High resolution live-video is converted into frame-by- frame image scans that provide 98% or greater image overlap. Subsequently, from the same extracted library, a photogrammetry model and high-resolution mesh is generated and stored with the initial archive. This model is broken down into an 18-point skeleton based on the CNNs training and understanding of which points are of value to the system.
IKIN’s photogrammetry – the science of making measurements from photographs – using AI and CNNs overcomes the challenges of accurate and repeatable scanning due to inconsistency of capture and poor rotational compensation.
“I’m impressed with how Cubara utilizes its CNN in order to achieve more accurate, error-free and repeatable scans,” TMC’s CEO Rich Tehrani.
But the aggressive scanning – called orthomosaic diagnostics – is only part of the story. Once you have the 3D image and data, it’s all about what you can do with it.
“It’s not the ability to scan, but the ability to detect change and use AI and analytics to predict change over time that’s really the big story,” said IKIN CEO Joe Ward.
Cubara’s diagnostic feature detection allows the smallest of changes in point cloud measurements to be detected, which correlate to changes over time on the scanned object. That includes changes not visible to the naked eye. With the highly accurate point cloud and 3D facial rendering, the cosmetics company was able to continue its testing and, in fact, is likely able to achieve better results than previously.
“IKIN’s Cubara scanning technology seems to be as advanced as their hologram tech which has impressed us continuously for many years,” noted Tehrani., who first got a preview of IKIN's tech back in 2019.
The cosmetics use case barely scratches the surface of what IKIN can do with Cubara. The implications are almost endless once you have the ability to create such high-density point clouds from images.
As Ward noted, “We really focused on scanning the face and we have created something unique that will be able to be positioned across multiple verticals. If you can scan a face, you can scan pretty much anything.”
That’s especially true when you factor is Cubara’s ability to create 3D models from 2D footage. In healthcare, for instance, providers can use the technology to examine wound conditions over time, where a CNN can be created for each wound to show they change over time.
The industrial sector can use it to better understand wear and damage to mechanical parts by creating temperature maps and understanding stress points in turbine blades, engines, and other equipment. Manufacturers and customers can much more accurately predict when products need to be replaced without having to wait for them to break, avoiding costly downtime and eliminating risk due to malfunction. Even clothing manufacturers can better understand how long their products and different materials will last under various conditions.
“We were just doing scanning for one company for clinical research,” said Taylor Scott, IKIN CTO. “We can use the same set of skills for everything on the planet, anywhere you want to see how patterns emerge or change happens over time.”
As the world continue to become increasingly digitized, the ability to take contextual data – especially from 2D tools – and create accurate, detailed 3D representations and then be able to extrapolate change data brings real-world scenarios into the virtual world to drive process and product improvements across industries.
“Cubara dimensional scanning is quite impressive,” Tehrani concluded. “It is the ideal bridge between the physical and digital worlds.”
Indeed, with the Metaverse poised to change how we interact in business and personal settings, the ability to create highly accurate, realistic representations will redefine how we perceive virtual settings. Instead of interacting in a cartoon-like virtual world, we may soon visit virtual worlds that much more closely resemble the real world.
That connection becomes evident when you think about combining IKIN’s technologies. Once you bring the data and 3D models from Cubara into IKIN’s Ryz or Arc holography platforms, we’re looking at e brand new world of gaming, training, education, testing and development, and so many other applications – even communications. In fact, with IKIN’s technology, we’re closer than you think to a real-world gesture-based interactivity – think Minority Report – where people can take 3D models and move them around naturally, without cumbersome, uncomfortable goggles.
“Our mission is to evolve the human-to-machine interface so it’s more comfortable and creates an emotional connection,” said Scott. “The key to that is not just a new interface, but a new way of capturing the world. A huge part of the Metaverse is going to be capturing the world in 3D.”
Edited by Erik Linask