TU-TESL

• 22

(photo source: brainchip) according to foreign media reports, at the neuroips 2019 conference to be held in Vancouver, Canada this month, brainchip will work with Tata consulting in India Services) to demonstrate gesture recognition technology using dynamic vision sensor (DVS) pulse input on the AKIDA neuromorphological technology platform. The AKIDA neuromorphological system on chip (SOC) technology will be used to recognize and classify the gestures of the participants. The dynamic vision sensor camera will detect the position of gesture and hand, and use the AKIDA platform for unsupervised real-time learning and classification. Compared with the traditional neural network (DNN), the pulse neural network (SNN) and AKIDA neural morphological chips have lower power consumption and require less data, brainchip said. AKIDA is a licensed IP technology that can be integrated into ASIC devices and used as SOC for deployment in monitoring devices, advanced driving assistance systems (ADAS), automobiles (AV), vision guided robots, drones, augmented reality and virtual reality (AR / VR), acoustic analysis and industrial Internet of things applications. AKIDA processes nerves and accesses memory at the edge, saves computing resources required by the CPU of the system host, and optimizes functions such as training, learning and reasoning. Roger Levinson, chief operating officer of brainchip, said: "recognizing simple gestures looks very simple at first, but in fact, in today's environment of human robot interaction, it is a revolutionary technological progress. We are very happy to be able to demonstrate this transformative technology in neurips, and we are happy to show how an AKIDA based platform can acquire images, train systems to identify what it has learned, seen, and more efficient and accurate than other solutions. " In November this year, the dynamic neural network of brainchip was patented in the United States, which is an important feature of its artificial intelligence processing chip AKIDA.