Nota’s on-device object detection model that enables smarter ADAS, shows state-of-the-art performance while running independently on small edge devices. During the PoC project with Intel, our model showed two times higher FPS(frames per second) compared to Intel’s OpenVINO.
Nota’s authentication solution allows man-free access control at a construction site or in a hazardous location without violating user’s privacy. This model can even work on low-end devices like Raspberry Pi with minimum latency (less than 0.5sec) and high accuracy level.
Nota’s real-time object segmentation model identifies the inventory level of the consumer goods on store shelves to give alerts when certain items are out of stock. By compressing the size of the model, we have reduced the computing resource and power used on the server, allowing the customer to run AI on larger number of cameras using the same amount of the CPU usage.
Nota Keyboard is a smartphone soft keyboard that reduces typo automatically by learning users’ input patterns itself without additional manipulation or user adaptation.