Even in difficult situations for recognition, real-time matching with millions of face databases is possible!
"SAFR(Ⓡ)" is a facial recognition software equipped with the world's highest level AI, trained on over 10 million facial data through deep learning. It can recognize faces even in real-world conditions (moving, obscured, wearing makeup, insufficient lighting, not facing forward). Additionally, the recognition accuracy of masked faces has dramatically improved due to enhanced Occlusion (foreign object removal) functionality. 【Features】 ■ Industry-leading high accuracy of 99.87% ■ Industry-leading speed of 117ms ■ High flexibility and scalability ■ Low cost for both implementation and operation ■ Minimal decrease in recognition accuracy based on race or skin color *For more details, please download the PDF or contact us.
Inquire About This Product
basic information
【Other Features】 - Ability to search for events and individuals from the activity list - Person detection possible from a part of the body (authentication not possible) - Real-time display of visitor numbers, attribute information, and summaries - Automatic adjustment of video contrast to improve object detection rates in low light - Collection of metadata such as gender, age, smile index, and gaze, etc. *For more details, please download the PDF or contact us.
Price range
Delivery Time
Applications/Examples of results
【Case Studies】 ■NTT docomo 'EasyPass(TM) powered by SAFR(TM)' ■Kolkata Police ■Indian Independence Day ■Itaú Bank ■Praia Grande Police, etc. *For more details, please download the PDF or contact us.
catalog（3）
Download All Catalogs
Company information
Our company leverages its achievements and innovations in the digital media business as a technological foundation to develop products that enrich and secure our lives using industry-leading AI and deep learning technologies. We offer tools such as 'SAFR' that are suitable for enhancing security, convenience, and attribute analysis. Please feel free to contact us if you have any inquiries.