An emotionally intelligent AR assistant is what this startup is building using Machine Learning capabilities.
By Srinath Srinivasan
Scanta is a research-first company which tries to bring to life what Augmented Reality (AR) has in store, using machine learning (ML) capabilities. Creating emotionally intelligent AR assistants is Scanta’s mission. With offices in San Francisco and Gurgaon, this startup wanted to build something that addressed day-to-day use cases with AR, says founder & CEO Chaitanya Hiremath. “We wanted to build something that wasn’t only for the businesses,” says Hiremath, talking about the unique route the business is taking today.
In 2017, Hiremath and his team started working on a Google tango project. It allowed 3D objects to be augmented through mobile phone cameras by mapping the environment. With this, Scanta created the world’s first basketball AR game post which it later moved to the Apple platform. By 2018, it started making AR avatars, around 110 of them, which would be crucial for content creation business. With this progress, Hiremath started pushing the boundary to create an entity which would enable two-way communication. Today, the entity can understand, react and respond in real time and animate an output based on the understanding, i.e., an emotionally intelligent AR assistant.
“One of the drawbacks when considering AR is that today it is mostly predefined animation. We want it to be much more than that in real time,” says Hiremath. He also demarcates various 3D imaging technologies from AR—for instance, holograms may not serve the same purpose as AR avatars all the time. “Our work is concerned with an asset library from which an animation can be chosen. We understand the sentiment of the users through facial features and signal processing of their voice. At the backend, we connect this understanding with our ML module which creates a conversation. Many startups are operating in the conversational tech space but what differentiates us is the visual appeal we create along with the audio,” explains Hiremath.
The avatar is a visual representation of the context in which the user is communicating to the system. Rather than collecting quantitative data, the company focuses on collecting qualitative data through a series of emotionally sound questions. Hiremath sees this technology scaling up to impact automobile industry, especially self-driving cars with heads up display (HUD), VR screens and virtual assistants. With Google launching AR maps in Pixel 3, this technology matches up to it—to have AR maps with avatars guiding the drivers. Other important industries where Hiremath thinks this can have applications are gaming, telecom, healthcare and manufacturing. He believes that when 5G picks up, two-way communication using avatars will be made seamless without any latency. Recently, the company shot an episode with Discovery Channel which has used Scanta’s technology to showcase the capabilities of AR and it is expected to go on air in October.
Scanta has seen investment of $700,000, a part of which was raised from a Munich based VC and a New York based VC. Recently, Scanta won the 2018 San Francisco Shark Tank Showcase, being the first Indian company to win the showcase.