Qualcomm Announces Next-Generation Processor Features to Deliver VR and AI

Tencent Digital (Su Yang) On October 17th, Beijing time, Qualcomm held the 4G/5G Summit in Hong Kong. Apart from fully describing the advantages of the Xiaolong 820 solution, the most eye-catching feature was the planning of new features for the next generation of products.

According to the official introduction, the next generation of Opteron processors will focus on VR virtual reality, security, ultra HD VoLTE video calls, scenario computing performance, and AI machine intelligence and machine learning. Among them, VR virtual reality, security and machine learning are As the top priority, MediaTek will face tremendous pressure.

It is reported that VR virtual reality will become the standard feature of the main products of Qualcomm's solution in 2017. The new solution will enhance the optimization of virtual reality graphics rendering and the support for built-in panoramic video capture features and external panoramic cameras.

In terms of security, the new processor will introduce biosensors based on the existing fingerprint identification, based on the independence of biological indicators, optimize the security of mobile devices, including iris recognition and facial recognition, and in addition to security, based on Long-term maturity of the market for biometrics authorization research, the new security program costs will also be optimized.

In terms of machine learning, Qualcomm will introduce solutions for products including mobile devices, drones, self-driving cars, smart cameras, etc., so that the machine can automatically adjust correction algorithms to form deep learning concepts such as image recognition. The new technology divides the graph into small pieces, and then accumulates them a little to produce the recognition result.

Qualcomm stated that machine learning is a multi-level process, and the industry already has related projects. The reason why Qualcomm is now advancing strongly is "because the current chip has a powerful data processing capability, and the market has a large number of data samples and algorithms with automatic correction. For other reasons."