Some innovations are driven by external objective factors. For instance, online shopping inherently demands highly detailed visualization of products, quickly revealing that a browser with limited resources struggles to showcase the realism of products and simulate their physical characteristics. However, first, the pandemic and then the drastic changes in climate conditions (such as this year's abnormal heatwave) pushed online shopping to the forefront of the market. Now, the capabilities of augmented reality have significantly enhanced this segment. For example, the ability to try on clothing, jewelry, and accessories virtually today almost matches reality in terms of realism. The ability to not only visualize clothing and accessories and try out different styles but also to create the illusion of a physical fitting elevates online shopping to a new level. It's no wonder that 92 percent of GenZ consumers, the first truly digital generation, express a desire to incorporate AR into their shopping experiences—the gap between physical and digital shopping experiences no longer exists.
Vladislav Bondarenko works precisely in this segment. He is a specialist in Software (Augmented Reality, Graphics, Machine Learning) and an Augmented Reality Engineer at try-on.io and Geenee.ar. Vladislav is a unique expert in the new augmented reality industry. He has worked at Delone, Innoscripta GmbH, and Clever Craft d.o.o for "Hewlett Packard." His portfolio includes working with AI face and body tracking, developing a physics engine for simulating clothing behavior in a browser, creating a highly realistic mechanism for displaying glasses, watches, clothing, and other consumer goods, and playing a key role in client projects for Shein, Sunglasshut (Rayban, Oakley, Versace), and Ripley's museum. Additionally, he regularly publishes educational articles explaining the possibilities of augmented reality.
Vladislav shared with us the simplicity of virtual fitting for the user and the colossal complexity of the developer's solutions that make this possible.
You focused on ML during your student years and won a national competition. How did your career develop from there?
In university, I participated in various ML competitions and projects. At 20, I won a nationwide competition in RF. During a week-long hackathon, I had to build an AI algorithm to identify and predict adverse DNA data based on characteristics from a table with around 900 columns. I managed to achieve the highest accuracy in the final table on my own. After university, I moved to California to work in the office for a year and a half. My interest in AR came to the forefront. Among the projects I worked on were various games for conferences with body control—the era of the metaverse and Non-fungible tokens, NFT, was just beginning, and everyone wanted something like that. I handled most of the software parts for Ripley's museum—where people could use screens with cameras to try on any costumes of their favorite Hollywood heroes. Later, our company decided to move towards interactive demonstration and virtual fitting of products.
Many of your developments are unique, with much time and effort invested in Research and Development because there were no ready-made tools or competitors to borrow from. How did your work in the augmented reality industry develop?
Our company profits from selling software products. We practically developed the technology from scratch, which today brings revenue. We have a small team of only six developers. I have been part of the Founding Team for almost four years, often leading projects. The idea was entirely different four years ago when I joined the startup. Over time, the company went through many changes and worked in various fields—from NFT to launching AR advertising. I conceived and executed some directions and technologies.
What is the purpose of augmented reality when it comes to specific applications, such as choosing jewelry online? What does virtual fitting entail from a developer's perspective?
The task is complex, requiring the combination of many technical details. For example, a user wants to try on earrings. They only need to press the "try-on" button and turn on the camera. For me, it involves fully forming a virtual reality: recognizing the ear's coordinates from the camera, tracking all movements in real-time, matching points from the user's camera with 3D space, positioning and animating the earrings, working with 3D graphics, and placing the earring in the correct place, accurately adhering to the laws of physics. The same applies to clothing: it is challenging to solve the problem of modeling fabric behavior, reproducing drapery and movements, and ensuring the natural interaction of the item with the user's body—their unique movements and speeds. This is where the power of physical simulations comes into play. Developing its own AR physics engine including—gravity, mass, collision, light, and shadow settings—allows virtual objects in the fitting scenario to react in real-time, providing a more realistic experience.
How does this work from the perspective of an ordinary user?
You visit a product website. For example, if you want to buy glasses. You need to try them on and get a complete impression of how they fit. Usually, fitting is done directly on the product page. When you click the "Try On Virtually" button, the camera turns on, and a fitting window opens right within the page. Everything happens through the camera in real-time, without uploading any files or photos. You can also change the color and size variations right there. The process gives you the most realistic illusion of a full-fledged fitting.
Is your method patented?
Yes, in the form of several patents. One of the last ones with my participation was received in May of this year. It is an original method for realistic product representation in a browser.
Today, this tool is used by major clients to increase sales and reduce returns. These include renowned companies like Shein, L'Oreal, Macy's, Sunglasshut, John Hardy, Ripley's, Sunski, and others. Why did your idea make such a splash? Wasn't it used before?
I certainly didn't invent it; the idea existed on the surface for quite some time. For example, it was used on Snapchat but was aimed at a different audience. However, no virtual fittings through the camera had been done in this particular way. There wasn't an online service helping the end customer get the best idea of a product, try it on without leaving home, and immediately make a purchase. The challenge was that, unlike platforms like Snapchat or Instagram, our product had to look better than theirs but in a browser, where the user's resources are much more limited compared to applications. We needed to achieve maximum product realism while ensuring smooth operation on most devices. Today, our products are implemented by large companies at the client level and are actively used. Retailers use AI-powered augmented reality try-ons to provide personalized shopping experiences based on the customer's purchase size and preferences. Most importantly, our efforts are scalable. Virtual fittings can be easily integrated into product pages on any e-commerce website, in-feed ads, as content for digital marketing campaigns, and as engaging touchpoints on the site, such as shopping windows or magic mirrors at live events.
