semiconductorsystem lab

|  Introduction  |  People  |   Publications  |  Project  |  Resource  |  Link  |  Contact us  |

Introduction

  Recently, object recognition has been widely used for various applications such as robot vision system, autonomous vehicle control and natural human-machine interfaces. The object recognition applications are characterized by complex and data-intensive computations, which make it difficult to achieve real-time performance even on today¡¯s state-of-the-art processors. Especially, real-time performance and low-power consumption are important requirements for embedded systems and programmability is also considered to deal with various recognition targets and algorithms. Such requirements of the mobile robot vision system motivate our research to design a real-time object recognition processor with low-power consumption.
  The purpose of our research is to develop a mobile robot vision system with real-time object recognition processor as a key component. Our research includes the following projects ranging from hardware issues to software issues regarding object recognition applications, all going to be merged inseparably.

- Bio-inspired attention-based object recognition algorithms
- Memory-centric network-on-chip for efficient data transactions of object recognition   applications
- Visual image processing memory specialized for object recognition
- Power efficient real-time object recognition processor with visual attention engine
- Real-time multi-object recognition processor with neural perception engine
- Development of low-cost vision platform for intelligent mobile robot