1~2 item / All 2 items
Displayed results
Added to bookmarks
Bookmarks listBookmark has been removed
Bookmarks listYou can't add any more bookmarks
By registering as a member, you can increase the number of bookmarks you can save and organize them with labels.
Free membership registration
Contact this company
Inquiry Form1~2 item / All 2 items

The "joint patent technology" and "eye detection technology (RGB version)" utilizing "deep learning" developed in collaboration with Panasonic and our company is a highly versatile elemental technology that allows for easy measurement of "blinking," "eye feature points," and "viewing distance (the distance from the camera to the eye)" using web cameras and device cameras on PCs. The detection uses a learning model based on deep learning, and we have compressed the learning model to achieve a processing speed of 50 to 60 ms per frame even with CPU processing (on our company's Windows machine with an 8th generation Core i7 CPU). We offer an SDK and an application, and the application supports CSV output and recording functions, making it capable of handling not only real-time verification. The application is available in both "camera input version" and "file input version."
Added to bookmarks
Bookmarks listBookmark has been removed
Bookmarks listYou can't add any more bookmarks
By registering as a member, you can increase the number of bookmarks you can save and organize them with labels.
Free membership registration
Built on technology jointly patented with Panasonic, our Liveness Detection (Anti-Spoofing) Technology is a biometric security core technology that prevents spoofing and hacking in systems such as facial recognition and eKYC (online identity verification). It works with visible-light cameras such as USB webcams and built-in PC cameras, and delivers high-accuracy determination using deep learning. It prevents spoofing through "Fake Face Detection," which can determine liveness from a single frame, and "Action Detection," which uses time-series data to detect specified actions such as blinking and changes in face orientation. By compressing the learning model, we achieve a processing speed of approximately 100 ms per frame even on CPU alone (measured on our Windows machine with an 8th-generation Core i7 CPU). We offer both an SDK and applications. The applications support CSV output and recording as standard, so they can be used beyond real-time verification. Two versions are available: a camera-input version and a file-input version.
Added to bookmarks
Bookmarks listBookmark has been removed
Bookmarks listYou can't add any more bookmarks
By registering as a member, you can increase the number of bookmarks you can save and organize them with labels.
Free membership registration