The NtechLab algorithm is able to detect an unlimited number of faces in an image, which makes it an ideal security solution. The NtechLab technology detects faces in complex conditions and is effective even when there is a significant change in light, posture, and face angle.
ON THE DETECTION
OF PEDESTRIANS AND CYCLIST
PERFORMANCE ON DETECTION
ACTIVITY IN EXTENDED VIDEO
#1 on Wild data set with FMR=0.001 and FNMR 0.1
NtechLab’s algorithm uses datasets of global scope. Billion-entry facial database searches are performed with the highest possible accuracy and speed.
Large database search speed:
- Set of 250 million images: less than 0.2 sec
- 500 million images: less than 0.3 sec
- 1 billion: less than 0.5 sec
The algorithm extracts the following attributes:
Our technology demonstrates the highest speed and accuracy in extracting attributes. The attribute data may be used in a wide variety of fields, including fraudster search, ad targeting, individualizing access control, and data enrichment. The accuracy of our emotion recognition algorithms is proven by NtechLab’s victory at the international EmotioNet Challenge 2017 contest.
NtechLab has developed it’s own 2D liveness technology based on passive detection that’s ideal for access control as well as for authorization through mobile apps purposes.
The key advantages are:
- No explicit actions required from visitors such as smile, blink or any other interactions with the system
- It works with existing equipment. No need to install specific devices, such as 3D cameras, thermal viewers and others
- Detection and verification takes less than 1 second
- Still images and video can be processed
- API-driven operation
- Its capabilities include:
- Accurate visitor counting when faces are not visible
- Silhouette-based trajectory detection
- Operates either as an add-on to a facial recognition system or as a standalone module
«The best face detection accuracy»
Outperformed Google and other competitors
«The best verification accuracy»
«The fastest identification speed»
«The best emotions recognition»
The 2nd best accuracy performance on detection activities in extended video