Panasonic has announced a facial recognition API. Panasonic has been a player in the facial recognition space for decades and has provided the technology that powers a number of real-world facial recognition use cases such as passport control and ticketless entry. However, until now, Panasonic has not distributed its technology to the developer community through an API.
Panasonic's approach to facial recognition uses two overarching methodologies that the developer community can rely on. First, Panasonic Feature Quantity Generation recognizes both the entire face and parts of the face by utilizing various deep learning technologies. Second, Panasonic uses proprietary algorithms to help limit errors by calculating a degree of similarity for matching faces. This combination helps expand Panasonic's usability as it continues to perform well at angles, poor lighting, or when a face is partially covered.
Panasonic is releasing the API as a microservice within its μSockets B2B IoT service. Currently, this service is only available in Japan. Within this service, Panasonic will offer both cloud-based and on-premise options to use the API. Panasonic will offer the API through a usage-based model based on the number of faces registered and number of recognitions performed. For more information, check out the facial recognition site.