China lags US in computer vision

China still lags the US in computer vision, despite its inroads in computing, research and its lead in data volume, a new study has found.

According to the inaugural Computer vision report by Hong Kong’s South China Morning Post (SCMP), the US remains the global leader in computer vision related hardware, talent and patents, surpassing China in the field.

In terms of talent alone, the US has 2,219 computer vision scholars, as compared to 190 in China, 362 in the UK, 262 in Germany and 259 in Canada.

That said, China is making up for its talent shortfall through paper representations at leading computer vision conferences. At this year’s Computer Vision and Pattern Recognition conference, it accounted for 39% of all submitted papers.

Computer vision currently makes up half of China’s artificial intelligence (AI) market. From 2016 to 2018, Chinese computer vision firms raised $4.5bn from venture capital, the highest globally.

SCMP’s 130-page report noted that China’s computer vision unicorns have made global headlines against the backdrop of rising US-China rivalry, adding that losing access to key component suppliers in the US in 2019 has not proven to be an existential threat to their survival.

Instead, China’s computer vision firms are grappling with major challenges at home. For example, much of their capabilities are currently deployed in government surveillance applications, but that market is becoming saturated, prompting a search for growth.

New growth areas include retail and marketing, where computer vision can pave the way for unmanned stores, which did not take off three years ago in China due to immature technologies, as well as autonomous vehicles for which Chinese technology giant Baidu has developed AI engines and applications.

“Having survived the US-China technological rivalry in 2019, computer vision unicorns in China face major challenges with an increasingly saturated market for surveillance applications,” said SCMP's technology editor John Artman.

“Searching for the next growth engine after security is a top priority for major computer vision companies in China and how they adapt will determine the evolution of the AI sector,” he added.

The SCMP report noted that China’s computer vision industry is also looking to rely less on underlying algorithms from the US, convincing its top talent to develop their ideas at home, ridding itself of its dependence on manually labelled data and addressing growing privacy concerns over facial recognition technology.

Under its national AI development blueprint announced in 2017, the Chinese government plans to grow the size of China’s AI industry to RMB 400bn (US$61.2bn) by 2025.

Besides China, Singapore has also outlined its AI ambition, starting with a national AI programme to build AI talent and start-ups, as well as boost AI adoption among businesses.

According to IDC, spending on AI in the Asia-Pacific region, excluding Japan, is expected to reach $4.6bn in 2021, with a compound annual growth rate of 72.9% between 2016 and 2021.