Past Projects

Jump to: navigation, search


High Efficiency Low Complexity Image Codec

  • Period : June 2013 ~ December 2013
  • Funded by Samsung Electronics
Samsung smartphone

In order to reduce power consumption in both the application processor (AP) and display driver IC (DDI), mobile image compress (MIC) will be applied in embedded devices such as the smartphones and tablet PCs.

The codec algorithm to be developed can reduce the power consumption for delivering screen data between AP and DDI. In addtion, the algorithm can be utilized to reduce the memory size of GRAM for storing previous frame.

Domestic Patent : 영상 처리 디바이스 및 방법, 대한민국 2014-0093296, Jul. 23, 2014. (applied)

Int. Proc. : “Parallel Implementation of Hybrid Vector Quantizer based Block Truncation Coding for Mobile Display Stream Compression,” International Symposium on Consumer Electronics, Jeju, Korea, Jun. 2014.

The Development of Automatic Merging Technique for 3D Face Avatar & Body Model

  • Period : July 2012 - January 2013
  • Funded by NRF (LINC) & Nanoinside
The flowchart of the system

Using the position information of landmarks (including eye corners, mouth corners, nose tip, and chin tip) detected from a scanned facial data, the facial data is aligned and merged to a predefined avatar body model in the system.

The merged 3D data can be fed into the 3D printer to generate a doll having the custumer's face.

Domestic Patent : 각 분할 스핀영상을 이용한 얼굴인식 방법, 대한민국 13-0010422, Jan. 30, 2013. (applied)

Int. Proc. : “Robust 3D Facial Landmark Detection using Improved Spin Images,” International Conference on Electronics, Information and Communication, Bali, Indonesia, Jan. 2013.

3D Image Retargeting

  • Period : May 2011 - April 2012
  • Funded by NRF (National Research Foundation of Korea)
Importance map obtained using 3D information & Retargeted images

The conventional image resizing method resizes images uniformly. As a result, some important objects in the scene can become unrecognizable.

Retargeting changes the dimension of the image non-uniformly. Important objects tend to remain, while background easily changes its dimension with unnoticeable deformation. To do this, retargeting requires image understanding to extract the amount of perceptual importance on each pixel/region.

The objective of the project is to devise an algorithm that represents the perceptual importance more accurately using 3D information.

Int. Journal : “Content-aware three-dimensional image retargeting for mobile devices,” Optical Engineering, vol. 51, no. 6, pp. 067014-1-10, Jun. 2012.

2018 Graduation Work

produced by Hakjong Kim and Woojin Choi.

Skin Analyzer is implemented.

In this work, Face skin is analyzed using GLCM to measure roughness and redness of the skin. Then, the system recommends cosmetics for the skin.

produced by Hyunsuk Kim, Younghyun Park and Minji Noh.

Solar Bench is implemented.

In this work, the solar cell panel makes shade for rest and generates electricity for providing several IoT services, such as wireless charging and dust forecasting.

2016 Graduation Work

produced by Chul-Woo Kim, Jong-Yong Kim, Tae-Doo Um.

A 3D model authoring system is implemented.

In this work, a 3D model is acquired with Kinect and rotating plate. The holographic projector displays the model that is manipulated with Leap Motion as input.

produced by Hakjong Joo, Seyoung Park, Songeun Kim, Dongmin Cho.

Cardio Mirror is implemented.

In this work, a mirror having webcam detects skin color variation on faces to measure the heart beat rate, and displays the results. The heart rate information is forwarded to user's mobile phone for logging and can be transferred to other persons if it is abnormal.

2015 Graduation Work

produced by Jae-Youn Jung, Dong-Kyun Woo, Sung-Jin Park.

In this work, Remote robot manipulation using Leap Motion was implemented.

A robot scans 3-D scenes around the robot with LIDAR sensor. The scene can be perceived in a remote site using Oculus Rift. Then, the robot arm is manipulated with Leap Motion as input.

2014 Graduation Work

produced by Byoung-Kwan Ko, Sang-Su Ahn.

A puzzle-solving robot was presented.

In this work, scattered puzzle pieces are recognized using a web camera, a robot places each piece to its right position and solves the puzzle.

2013 Graduation Work

produced by Myoung-June Kim, Hyoung-Chul Ahn, Jin-Hyun Park, So-Hee Park

Motion-Capture Drum Game
Uisng Kinect, gamers' motion is captured for playing drums.

produced by Hyoung-Soon Jeon, June-Soo Choi, Young-Jae Sohn

Mobile Guitar Tutor
An Android device instructs guitar chords through Bluetooth.
The chords are indicated using LEDs inlaid in guitar neck.

2012 Graduation Work

by Hyosang Jo, Jaehyuk Jang, Youjin Park, and Woochul Shin

Whos here now:   Members 0   Guests 1   Bots & Crawlers 0
Personal tools