Member Spotlight: Masaki Hilaga, Morpho

Silicon Foundry’s expanding network of Corporate Members are driving transformation at some of the world’s most respected global brands — from leaders in mobility, mining and telecom, to beauty, retail, energy, financial services and beyond.

Masaki Hilaga is the Co-Founder and CEO of Morpho, a developer of image processing technologies based on his work as a researcher at the University of Tokyo. Founded in 2004, Morpho’s technology has been used in over 3 billion devices worldwide. The company was listed on the Tokyo Stock Exchange in 2011 and has subsequently begun expanding its business to international markets.

Tell us a little about Morpho and yourself.

Morpho actually began when I was a PhD student at the University of Tokyo studying computer graphics and computer vision. I worked as an engineer for about a year and a half after I finished the PhD program, but couldn’t shake the feeling that I wanted to do something myself. It was around that time when I met Chiaki Konagai (Morpho’s co-founder), who had already achieved a career as a director of a public-listed company on the Tokyo Stock Exchange.

I told him about some of my thoughts and he invited me to start a company together, so we ended up incorporating Morpho with the support of the University of Tokyo, which had a startup incubation program. At the time, we raised 50 million JPY (~$480,000 USD) from the University’s venture capital fund and launched the company. Morpho was the first company founded through the University of Tokyo to go public.

Since Morpho’s establishment in 2004, our primary focus has been the development of image processing and computer vision solutions for a broad range of use cases. Traditionally, the largest component of our business has been mobile image processing solutions for smartphone manufacturers and chipmakers — classic examples of this include image stabilization for stills and video and panorama stitching. We were actually one of the first to commercially launch a panorama stitching feature for mobile phones in 2006 — close to 5 years before Apple brought that feature to the iPhone in 2011.

For a brief overview of Morpho, can you describe your core products and technologies, as well as your main business and investment focus areas?

More recently, Morpho has been developing edge AI solutions that utilize deep learning algorithms and machine learning for applications ranging from smartphones and autonomous vehicles to IoT and smart city applications. Examples of this include semantic filtering technology for pixel-level object recognition, congestion analysis and crowd counting for social distancing, and medical diagnostics. To that end, we’ve been actively engaging with a number of different external partners around the world to co-develop solutions. International expansion is also a key priority for us at this time; we’ve recently opened offices and expanded teams in Taiwan, the US, China, and South Korea. We’ve also acquired a AI developer named Top Data Science in Finland. We think opportunities in this space are endless — and we’re just getting started.

Given the recent formal partnerships Morpho has formed with DENSO and SRL, how would you describe Morpho’s business development goals via these strategic alliances?

Given our core focus on effective research and development since our establishment, our open innovation policy is a key facet of our strategy — we’ve been collaborating with a number of external partners to offer products and solutions that can drive significant positive market and social impact.

In terms of our specific engagement with DENSO, we are currently engaged in the joint development of ADAS and autonomous mobility solutions by utilizing image and vehicle control data that they’ve accumulated over the years. For our engagement with SRL, which is known for for COVID-19 PCR testing in Japan, we have leveraged their enormous library of data to improve the efficiency of medical diagnostics, such as the detection of chromosomal abnormalities, with our AI technology. Many of these projects would have been otherwise impossible for us to achieve by ourselves.

With the newly established Morpho subsidiaries in South Korea, Taiwan, the US, China, and Top Data Science in Finland, can you talk about your strategic goals for international expansion?

Over the course of the last few years, we have been working to internationalize our operations and are aiming to establish specialized teams — including local R&D and customer service functions — in a number of strategic markets. We believe that growing our global presence is one of the best ways to gain a more intimate understanding of our customers and meet local and regional market needs. South Korea, Taiwan, and China in particular are growth markets that we think are highly promising.

To elaborate on our acquisition of Top Data Science, a Helsinki-based AI/ML technology developer, we believe that it will strengthen our technology portfolio and help us establish a foothold into European markets. Ultimately, Morpho’s vision is to establish a network of mutually supporting offices that are able leverage each other’s solutions and know-how — our Tokyo office won’t necessarily be where all of our most important work is done.

In terms of our US operations, we’ve been exploring a number of areas with Qualcomm, one of our key strategic partners. In addition to our primary work with them on smartphone image processing, we’ve also been looking at applications of our technology for security cameras, drones, and head-mounted displays. Recently, we were able to become a software partner of NVIDIA Metropolis program, an initiative to develop an IoT ecosystem relating to smart city technologies on NVIDIA’s AI platform. Given the role that the US and Silicon Valley play in pushing technological boundaries forward, we’re eager to engage with potential partners here to co-develop solutions together.

What is your expectation and outlook for the future of AI technology and solutions?

The introduction of deep learning was particularly momentous because of its large range of applications — you can use it for image recognition, audio recognition, natural language processing, and other uses — and achieve great results without being an expert in the field. This is something that wasn’t possible before.

This is an amazing development, but something that has significant implications for us as a company. The lowering of barriers into image recognition solutions means that more non-experts are now able to enter the space and accelerate technological advancement, but also entails the intensification of competition for us as a company. That being said, from the perspective of the average consumer, all of this is fantastic because the world becomes more convenient as a result. We’ll be doing our best to contribute to accelerating the development of useful technologies, keep pace with the best players in the space, and work to accumulate specific industry knowledge and experience to create products with strong value propositions.

Ultimately, AI is one tool amongst many — I really think that the most important thing is what it’s used for. One trend that’s clearly picking up steam lately in our core business is an image enhancement software such as AI-based segmentation and pixel filtering. Users will be able to capture professional-level images through features that automatically recognize subjects (e.g., human hair, skin, green, sky, water) and image optimization with a single touch of the shutter button on a smartphone. We were able to win Best AI Software/Algorithm Award from Edge AI Vision Alliance for this product in July this year.