Human Ethnicity Recognition 07.10.17

Face-Recognition Site Reveals Shades of Gray in Human Ethnicity

Asian, black, Latino or white? In the ethnic melting pot that is the United States today, these overgeneralized racial classifications are becoming increasingly inadequate when it comes to describing someone’s background. To address the shortcomings of these blanket classifications, a company called Kairos is using face-recognition software to estimate the percentage of a person’s ethnic makeup based on his or her facial features.

Kairos Diversity Recognition is a face-recognition web application that shows the diversity and nuances of ethnicity. Simply by loading a picture of a face to the site, users can see an estimate of the subject’s percentage of ethnic background.

For example, an analysis of the visage of actress Michelle Rodriguez indicates she is 50 percent Hispanic, 38 percent white, 5 percent black, 3 percent other and 2 percent Asian. In contrast, fellow actress Lucy Liu is estimated to be 99 percent Asian.

The company estimates that more than 7 million facial images have been loaded to its website so far.

Kairos describes its technology as a combination of computer vision and machine learning. The company uses proprietary face analysis and machine learning algorithms which are undergoing constant improvement.

Kairos’s technology is also available as a cloud-based application programming interface, which developers can use to incorporate face-recognition into their software with just a few lines of code, according to the company. The company said brands can use the technology to track critical demographic data regarding their audiences.

Furthermore, Kairos offers a software development kit that allows companies to integrate the capability to gather actionable, real-time data about people right inside their products.

Beyond ethnicity, Kairos says it can gather a wide of other data types from its face-recognition technology. The company said it can conduct face detection, identification, and verification. It also can detect emotions, age and gender.

Moreover, the software can measure how much people are paying attention and determine user sentiment.

Kairos acknowledges that establishing DNA genealogy requires biological testing and mapping a family tree to determine ancestry. However, the company said its approach provides a faster route to gathering demographic data.

“…what sets our Ethnicity Detection apart from both DNA genealogy and ancestry is time,” the company notes on its website. “From a single picture, in seconds, you get a highly accurate breakdown of your ethnic makeup. No digging through records, or sending DNA through the mail.”

Nirel Marofsky is project analyst for the cognitive engine and application ecosystem at Veritone. She acts as a liaison to strategic partners, integrating developers and their capabilities into the Veritone Platform. Learn more about our platform and join the Veritone developer ecosystem today.