Home News Details

Hi3559A’s Neural Network Inference Engine drives the development of Multi-Object Categorization and Identification System

In recent years, AI has become to be the most popular topic not only in the industry but also in our daily lives. How best to utilize the power of AI to transform what we do. HiSilicon’s newly announced Hi3559A, is the first step in utilizing the power of AI, for vision capture applications.

Hi3559A is based on a scalable computing architecture, & Smart Vision Platform (SVP), which is able to support high performance vision processing with trillion operations on the efficient heterogeneous computing platform. Standard and custom functions can be accelerated through vision DSP and IVE. As the key module in SVP, HiSilicon’s xNNIE engine is superior to others, where Deep Learning Neural Networks are fully supported by our dedicated hardware accelerations. With the xNNIE engine, user applications can be easily implemented in Hi3559A, such as object categorization (motor vehicle, non-motorized vehicle, and pedestrian detection) and attribute recognition (vehicle type, color, and license plate recognition).

HiSilicon has successfully demonstrated the real object categorization capabilities using the xNNIE AI technology of the Hi3559A at several recent events. We welcome you to join us during any of our upcoming 2018 events and discuss more about the future and exciting possibilities of AI.

Please visit our website for more event information.

HiSilicon is proud to help our clients make leading class products through the use of our state-of-art technologies.

Multi-object Categorization and Identification System

Key Features

  • High performance vision processing

  • Scalable computing architecture

  • Hardware acceleration of Deep Learning Neural Networks on the dedicated engine

  • Generic object categorization and detection using Convolutional Neural Networks

  • Object Detection: sofa, desk, chair, bus, car, airplane, bicycle, motor, cat, dog, bird etc.

  • Attribute recognition: vehicle type, color, and license plate recognition etc.