Sensors on Delay Compared with Algorithms

This post is also available in: עברית (Hebrew)

A new development that may boost facial recognition technology enables secure biometric authentication for a variety of banking, e-government and enterprise applications. However, sensor quality in the biometric field still needs improvement.
Lithuanian biometrics firm Neurotechnology announced the release of their new Face Verification SDK (Software Development Kit). The technology aims to provide an easy-to-use authentication platform for implementing facial recognition for payment authentication and logins for online services.
The technology is a subset of the company’s VeriLook SDK, which it advertises as being capable of simultaneous multiple face processing in live video, as well as gender classification, age detection, emotion recognition, and facial feature points and attributes detection.
In a press release, the company notes, “The face image checks may be used by the integrators to ensure the quality of the captured face image and used in combination with liveness detection to reinforce anti-spoofing measures.” The SDK also permits face verification to be performed offline, according to techrepublic.com.
While advances in recognition algorithms are important, improvements are more pressing on the sensor side to provide higher quality input for the algorithms to analyze. Sony’s sensor head Satoshi Yoshihara told bloomberg.com that 3D camera sensors with advanced depth sensing are coming in 2019. Sony’s depth sensing method relies on measuring the time it takes for invisible laser pulses to travel to the target and back to the handset.
Sony already controls roughly half of the camera chip market. In addition to being used in Sony’s own Xperia smartphone line, the chips are also used in iPhones, in Samsung’s flagship Galaxy S and Note series, and in Google’s Pixel phones, in addition to other Android smartphones.