Jongju Shin

[Information]

Vision Intelligence Team Lead,

Kakao Enterprise,

3rd floor, H-Square, 235 Pangyoyeok-ro, Bundang-gu, Seongnam-si, Gyeonggi-do, 13494, Korea

E-mail : isaac.j.shin@gmail.com,

             jjshin@postech.ac.kr,

             isaac.giant@kakaoenterprise.com

[Educational Background]

1999.03~2007.02 B.S., Dept. of Information & Computer Engineering, Ajou University, Suwon, South Korea.

2007.02~2015.02, Ph. D., Dept. of Computer Science & Engineering, POSTECH, Pohang, South Korea.

2011.09~2012.08 Visiting Student, Robotics Institute, Carnegie Mellon University, Pittsburgh, PA, US

2015.01~2015.05, Post-Doc., Dept. of Computer Science & Engineering, POSTECH, Pohang, South Korea.

[Work Experience]

2002.04~2003.04 EBI solution (closed)

2003.05~2004.05 Goldenpond (closed)

2004.05~2006.02 Unidocs

2015.06~2019.12, Research Engineer, Kakao Corp., South Korea.

2019.12~2021.07, Research Engineer, Kakao Enterprise, South Korea.

2021.08~current, Vision Intelligence Team Lead, Kakao Enterprise, South Korea.

[Research Interest]

Face related problems - detection, alignment(including tracking), recognition, attribute(expression, gender, age, etc) estimation.

Face Recognition (Kakao Enterprise, Kakao)

Interview

  


Facial Feature Tracking in Mobile Device (Kakao, closed service)


Daum roadview blurring

To blur pedestrians and car plates, we developed face detection, and car plat detection algorithms. (link)


Celeb Image Ranking Model

Rank images according to the qualities such as the number of faces.


Facial Feature Detection and Tracking (POSTECH)

Hybrid approach to combine local and global methods.

Facial Feature Detection (CMU)

All pose face alignment robust to occlusion (link)

Based on multiple shape models and RANSAC framework, we can align face robust to occlusion.

This work is done with Takeo Kanade and Myung-Cheol Roh in CMU.

Facial Feature Tracking (POSTECH)

Based on Active Appearance Models, we add global and local template matching, template update, and motion estimation.

We compared our algorithm with others. The result on the right bottom is ours.

In detail, We show the result with big head with mouth motion.

Facial Feature Tracking and Head Pose Estimation (POSTECH)

We combined cylindrical head model and facial feature tracking(based on Inverse Compositional image alignment algorithm with face shape model). The technique of facial feature tracking here is used for lip reading system in my journal.

[Publications]

International Journal


Technical Report


International Conference


Domestic Journal


Domestic Conference

[Projects]


[Patents]


- "그룹 기반 얼굴 인식 방법 및 장치",

  김용현, 박원표, 노명철, 신종주, 출원번호 10-2020-0059683

- "얼굴 특징점 추출 방법 및 이를 수행하는 장치",

  김대진, 신종주, 출원번호 10-2014-0054087, 등록번호 101626050000

- "평균 얼굴 패치 및 메시 기반 나이 추정 시스템 및 방법",

  전승선, 홍윤정, 김대진, 신종주, 출원번호 10-2014-0110960

-  "립리딩과 음성 인식 통합 멀티모달 인터페이스 시스템",

  김대진,이진,신종주,이진석,김대희, 한국, 등록번호 10-1092820 (20111205)

- "정면 얼굴 영상 생성 장치 및 방법",

  김대진,신종주,김정태,양승준,임성용,이한규,한국,출원번호 10-2010-0126120 (2010.12.10)

- "사용자 맞춤형 표정 인식 방법 및 장치",

  김대진,신종주,천영재,한국,등록번호 10-0988326 (20101011)

- "아이 컨택 시스템을 위한 얼굴 특징점 추출 기반 정면 얼굴 합성",

  김대진,신종주,김정태,양승준,임성용,이한규,한국,출원번호 10-2010-0007031 (2010.01.26)

- "COMBINED LIP READING AND VOICE RECOGNITION MULTIMODAL INTERFACE SYSTEM",

  Dae Hee Kim, Dai-Jin Kim, Jin Lee, Jong-Ju Shin, Jin-Seok Lee, US, 20110071830 (2011-03-24)

- "Method for correcting user's gaze direction in image, machine-readable storage medium and communication terminal"

  Byung-Jun Kim, Dai-Jin Kim, Jong-Ju Shin, In-Ho Choi, Tae-Hwa Hong, US, 20140016871 (2013-02--01)