Deep Learning & Computer Vision





   
Sungeun Hong
,
    Research Scientist & Photographer
    Email: csehong@sktbrain.com
    Office: T-Brain, SKT Tower, Eulji-ro 65, Jung-gu, Seoul
    Mobile: (+82)10-8711-6338.

    CV · Google Scholar · LinkedIn · GitHub · PhotoBlog  
.
 About me
    I am a research scientist at T-Brain of SK telecom, applying deep learning to the analysis of multimodality across a variety of research fields. 
    .
 Education
    KAIST, Ph.D., Computer Science, 2012 - 2018
    KAIST, M.S., Computer Science, 2010 - 2012
    Hanyang, B.S., Computer Engineering, 2006 - 2010  


 Research Interests
    Domain Adaptation  ·  Multimodal learning
    Universal Representation  ·  Face-related recognition
    Texture recognition  ·  Video scene recognition



Publications  

Doubly Nested Network for Resource-Efficient Inference

Jaehong Kim, Sungeun Hong, Yongseok Choi, and Jiwon Kim
<Submission> Jun 2018, arxiv
[PDF]
.
S
cale-Varying Triplet Ranking with Classification Loss for Facial Age Estimation
Woobin Im, Sungeun Hong, Sung-Eui Yoon, and Hyun S. Yang
<Accepted> Dec 2018, Asian Conference on Computer Vision (ACCV '18)
[PDF
.
Cascaded Pyramid Network for 3D Human Pose Estimation Challenge

Sungeun Hong, Wonjin Jung, Ilsang Woo, and Seung Wook Kim
Sep 2018, European Conference on Computer Vision Workshop (ECCV Workshop '18)
[PDF
.
Bilinear Attention Networks for VizWiz Challenge
Jin-Hwa Kim, Yongseok Choi, Sungeun Hong Jaehyun Jun, and Byoung-Tak Zhang
Sep 2018, European Conference on Computer Vision Workshop (ECCV Workshop '18)
Ranked 2nd in VizWiz Grand challenge, ECCV Workshop
[PDF]  [Slides [Poster]
 .
CBVMR: Content-Based Video–Music Retrieval Using Soft Intra-Modal Structure Constraint
Sungeun Hong, Woobin Im, and Hyun S. Yang
Spotlight, Jun 2018, International Conference on Multimedia Retrieval (ICMR '18), ACM 
[PDF]  [Slides [Poster]  [Demo]


SSPP-DAN: Deep Domain Adaptation Network for Face Recognition with Single Sample Per Person
Sungeun Hong, Woobin Im, Jongbin Ryu, and Hyun S. Yang
Oral, Sep 2017, International Conference on Image Processing (ICIP '17), IEEE
[PDF [Slides [Dataset]
  ..
D3: Recognizing Dynamic Scenes with Deep Dual Descriptor based on Key Frames and Key Segments
Sungeun Hong, Jongbin Ryu, Woobin Im, and Hyun S. Yang
Jan 2018, Neurocomputing, Elsevier
[PDF [Slides]  [Dataset] 
  ..
Discriminant Texture Classification based on Random Difference Forest
Jongbin Ryu, Sungeun Hong, and Hyun S. Yang
<Under Review> Pattern Recognition, Elsevier
[PDF]
  .
Not all frames are equal: aggregating salient features for dynamic texture classification
Sungeun Hong, Jongbin Ryu, and Hyun S. Yang
Jan 2018, Multidimensional Systems and Signal Processing, Springer
[PDF [Slides] 


Deep CNN-based Person Identification using Facial and Clothing Features
Sungeun Hong, Woobin Im, Junwoo Park, and Hyun S. Yang
Jun 2016, Summer General Conference '16, IEEK
[PDF]  [Slides] 
 
Sorted consecutive local binary pattern for texture classification
Jongbin Ryu, Sungeun Hong, and Hyun S. Yang
Jul 2015, IEEE Transactions on Image Processing (TIP), IEEE
[PDF]
.
Multi-view based robust face recognition
Byungok Han, Sungeun Hong, and Hyun S. Yang
Jun 2015, Summer General Conference '15, IEEK
[PDF]
  .
Recursive Bayesian fire recognition using greedy margin-maximizing clustering
Sujung Bae, Sungeun Hong, Yeongjae Choi, and Hyun S. Yang
Nov 2013, Machine Vision and Applications (MVA), Springer
[PDF]

  .
3D Keypoint Tracking Based on Hybrid Flow Computation for Human Motion Analysis
Sujung Bae, Sungeun Hong, and Hyun S. Yang
May 2013, Machine Vision and Applications (MVA '13), IAPR
[PDF]

  .
A License Plate Recognition System Robust to Vehicle Location and Viewing Angle
Sungeun Hong, Sungsoo Hwang, and Seongdae Kim
Dec 2012, Journal of the Institute of Electronics and Information Engineers, IEEK
[PDF]

  .
Edge-preserving Compression of License Plate Image using Its Bimodal Characteristic
Sungsoo Hwang, Sungeun Hong, and Seongdae Kim
Feb 2012, International Conference on Electronics, Information, and Communication (ICEIC '12), IEEE
[PDF]
  .
Edge-based License Plate Detection Method with little consideration of threshold determination
Sungeun Hong, Sungsoo Hwang, and Seongdae Kim
Nov 2011,  publication description Fall General Conference '11, IEEK
[PDF]
  .
License Plate Detection Based on Cascaded Three Distinct Features
Sungeun Hong, Sungsoo Hwang, Jisung Yoo, and Seongdae Kim
Jun 2011,  International Technical Conference on Circuits/Systems, Computers and Communications (ITC-CSCC'11), IEEK
[PDF]



Patent

  • Apparatus and Method for Providing Heterogeneous Contents Recommendation Model, Pending, Jul 18, 2017,  kr 10-2017-0090765
    [PDF]
  • Apparatus and Method for Providing Object Recognition Model based on Domain Adaptation, Pending, Jul 3, 2017,  kr 10-2017-0084235
    [PDF]
  • Partial Face Based Person Identification Across Poses, Issued, Jul 27, 2018, 10-1884874
    [PDF]
  • System and Method for Compressing Vehicle Image for Detection of License Plate, Issued, Nov 6, 2013, kr 10-1328533
    [PDF]
  • A License Plate Recognition System Under Various Environmental Conditions, Issued, Oct 29, 2012, C-2012-021243
    [PDF]
.
 Academic Activities
  • Reviewer
    • IEEE Transactions on Image Processing (TIP), 2018 
    • IEEE Transactions on Multimedia (TMM), 2018 
    • ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM), 2018 
    • IEEE Transactions on Image Processing (TIP), 2017 
    • IEEE Signal Processing Letters (SPL), 2016 Elsevier Neurocomputing, 2015 
    • IEEE International Conference on Virtual Systems and Multimedia (VSMM), 2014 
    • Digital Heritage International Congress, 2013 IEEE Virtual Reality (VR), 2011 
    • ACM SIGGRAPH VRCAI, 2010 
  • Lecturer
    • Course: ‘Overview of Domain Adaptation’ at SK AI Course, May 2018 – Jun 2018 
    • Tech Talk: ‘Deep Domain Adaptation Network for Face Recognition with Single Sample Per Person’ at Naver Tech Talk, Apr 2018 
    • Course: ‘Introduction to Programming using Python’ at KAIST, Sep 2014 – Aug 2017
  • Etc. 
    • Organizer and Manager at TEDx KAIST (tedxkaist.com), Aug 2012 – Mar 2013

 
 Honors & Awards
  • The first prize in SK telecom Tech Expert Board (TEB) excellent work awards, Oct 2018
    • Main tasks: 1) SK AI course lecturer 2) Naver tech talk 3) ranked 2nd in VizWiz Challenge
  • Ranked 2nd in VizWiz Grand Challenge: Answering Visual Questions from Blind People (ECCV Workshop) , Sep 2018
    • Title:  Bilinear Attention Networks for VizWiz Challenge
  • Naver Ph.D. Fellowship, Sep 2017
    • Ph.D. student who has published outstanding papers as the first author in an international top-tier conferences or journals
  • Outstanding Teaching Assistant Award, KAIST,  Feb 2017
    • Role: Python programming instructor for undergraduates  
  • LG Electronics Paper Award at IEEK Conference, IEEK, Jun 2016
    • Title: Deep CNN-based Person Identification using Facial and Clothing Features 
  • Winner of Idea Contest for R&D Development Special Zone Support Division, Apr 2012
    • Title: Research complex childcare system
    • Award: Special Award
  • Winner of Future Idea Competition, KOLON-KAIST Life Style Innovation Center, Oct 2011
    • Title: Realistic sports broadcasting through sharing senses between athletes and viewers
    • Award: Grand Prize (10,000$)

 Project
  • Visual Question Answering for Visually Impaired People
    •  July 2018 – Present, funded by AI Research Center, SKT 
  • Neural Architecture Search for Large Scale Image Classification
    • Jan 2018 – May 2018, funded by AI Research Center, SKT
  • Multi-view Face Recognition based on Deep Learning
    • Aug 2012 – May 2017, funded by Electronics and Telecommunications Research Institute (ETRI)
    • "The development of automatic user information (identification, behavior, location) extraction and recognition technology based on perception sensor network under real environment for intelligent robot"
  • Person Identification using Facial Attributes (Race, Gender, and Age)
    • Nov 2013 – October 2015, funded by National Research Foundation (NRF)
    • "Fast, accurate, robust identity recognition technology in a large database based on facial attribute information"
  • Spontaneous Facial Expression Recognition
    • Aug 2012 – Apr 2015, funded by National Research Foundation (NRF)
    • "Speech-image integration model based on brain information processing mechanism and its application to implicit Intention Recognition"
  • Vehicle License Plate Detection and Recognition in Real Environment
    • Aug 2010 – Feb 2012, funded by AddPac Technology
    • "A License Plate Recognition System Robust to Vehicle Location and Viewing Angle in Real Environment"