Smart Camera Identifies Faces Better Than Human Eye

This post is also available in: עברית (Hebrew)

A new HD smart camera boasts serious artificial intelligence capabilities and can identify faces with an accuracy of up to 99.7 percent. Its ability to capture and identify images simultaneously is a major breakthrough in the field of embedded A.I. in security and commercial applications. Startup Horizon Robotics claims the new HR-IPC2143 smart camera has applications in public security for the real-time detection of potential suspects. In the commercial sector, it can identify important customers.

The low-power low-cost camera utilizes cutting-edge artificial intelligence technology. The camera leverages recognition capabilities alongside its local capture to give users facial capture, feature extraction, and facial feature comparison based on deep-learning algorithms, according to digitaltrends.com.

The camera’s Sunrise embedded A.I. computer vision processor boasts the Brain Processing Unit (BPU) architecture. This proprietary A.I. processor architecture claims to combine A.I. algorithms with hardware design, and allows the relatively small camera to still perform at its peak.

This camera can quickly turn raw video data into actionable insights, with multi-target location detection, multi-target tracking, multi-target recognition, and pedestrian attribute classification. Thanks to its 50,000 feature library, the camera is able to support high-performance facial recognition, as well as process videos at 1920x1080p at 30 frames per second in real time. It also allows the camera to detect, track, and identify up to 200 objects simultaneously, all in a single frame.

The system can recognize faces even in crowded environments, like packed hallways, entries and exits, and large events. Or, if a store owner wants to identify VIP shoppers as they enter the premises to ensure that they receive special treatment, the smart camera can be leveraged in this way as well.