컨텐츠 바로가기 영역
주메뉴로 바로가기
본문으로 바로가기
A better IT world for all,
Center for Integrated Smart Sensors is making it with creative SMART IT convergences system technologies, we can make a better
where everyone's dreams come true.
About CISS Research Team

Research Team

Organization Board Auditor Director Research Team Administrative Team Technology Commercialization Team Evaluation Committee

3-4-1 Project Director

이미지
Chong-Min Kyung
    • Affiliation KAIST (Electrical Engineering) / CISS
    • Tel +82-42-350-3423
  • Research Topic Low-power and low-cost depth estimation camera system, smart camera hardware platform, smart black-box system

Project : Low-power Mobile 3D Camera System Design

  • Purpose

    In this project, we design and implement low-power and low-cost smart camera system, which has depth extraction capability. - Implementation of smart camera hardware platform including ISP(Image Signal Processor) and DMP(Depth Map Processor) - Development of black-box system which has long life-time as smart camera application

  • Contents

    Smart camera have ability to process data and provide various information users wanted, so that it can be utilized in diverse application area. In order to maximize these abilities of the smart camera, it must be operated with minimum power consumption with result of optimization considering rate and distortion for long life-time and utilization in mobile devices. For development of low-power and low-cost smart camera system, we are focusing on following topics. • Smart camera system design: Most surveillance cameras today consume too much energy, or most low-power cameras today are not smart enough. We implement low-power, low cost and reliable camera system which can be used and applied in practical situation. It includes development of system architecture, design of system architecture, event detection algorithm, image signal processor and CMOS image sensor. • Development of depth estimation algorithm: Obtaining depth information from camera consumes heavy resource, because its algorithm requires long computing time, high memory usage and high hardware complexity. We develop fast and accurate depth estimation algorithms which can extract depth information from blur based image using single camera. It requires low-power, low hardware complexity and it should be performed in real-time with high accuracy.

  • Expected Contribution

    The smart camera system will be widely utilized in civilian sector, such as automotive auxiliary sight device, facility maintenance and medical bio-signal sensing. As the smart camera is manufactured at low price, its application area is widely expanded so that they will make immense impact on diverse industry. Especially ultra-low power depth estimation camera will create whole new application area, and will lead market growth. This will accelerate research and development of novel security surveillance camera, and play a pivotal role in solving social issue which recently came to the fore worldwide.

List