The US Patent and Trademark Bureau has approved Apple’s patent application for a scanning system that can identify a person even without a clear image of the face in the frame.
On November 26, the patent describes a system that utilizes person recognition technology as the primary method of identification. When the image quality is inadequate, the system uses additional body characteristics to confirm the identity.
The technology works by initially recognizing the face in a video stream and then storing additional data about the person’s body, such as gait and clothing features. This data is utilized in scenarios where a clear face cannot be identified in a new video recording, possibly due to poor lighting. By comparing body movements and other parameters with existing records, the system determines if the person matches a previously saved profile.
“Based on the comparison, the device can send a notification indicating whether the identity of the second person matches that of the first,” the application explained. “This technology enables the device to identify a person without requiring a clear or high-quality image of the face.”
The system creates a “movement imprint,” which captures and retains unique body movement characteristics based on collected data.
According to 9TO5MAC, Apple refers to this innovation as “personality recognition using body characteristics associated with the face.” The technology aims to address situations where “a video camera is unable to perform facial recognition in a specific video stream.”
The patent application suggests that the system can be deployed not only in homes but also in office buildings, parking lots, public parks, and warehouses.
“It should be noted that while the implementation options of the present invention are primarily focused on performing identity recognition at home, they are not limited to this scenario,” the document stated.
However, not all technologies patented by the company are necessarily integrated into its products.