Description

Deep Vision (now Kinara) is an embedded deep learning solution designed for application developers. It enables users to detect signatures directly from the raw data, leverage layers of convolutional and recurrent neural networks to predict data, experiment with the relationship between accuracy, sample rate, window size, and computational intensity, reuse of training objects for validation, as well as embed computer visions with devices like drones, robots, and IoT cameras. Features include CNNs, traditional computer vision algorithms, scalable processor architecture, a stack of complementary software, and real-time processing.