Jinsoo Leo Choi

Research Interests

• Computer Vision / Deep Learning

• Video frame interpolation / Video stabilization / Super resolution

• Vision & language

Education

Korea Advanced Institute of Science and Technology (KAIST)

PhD in Electrical Engineering (Sep. 2015 - Dec. 2019)
Advisor: In So Kweon

MSc in Electrical Engineering (Sep. 2013 - Aug. 2015)
Thesis: Human Attention Estimation for Natural Scene Images via Gaze and Saliency Models
Advisor: In So Kweon

BSc in Electrical Engineering (Sep. 2009 - Aug. 2013)
Full scholarship

Publications

Deep Iterative Frame Interpolation for Full-frame Video Stabilization,
Jinsoo Choi, and I.S. Kweon,

ACM Transactions on Graphics (TOG/SIGGRAPH Asia), Nov. 2019.
[Paper] [Video]

Image Captioning with Very Scarce Supervised Data: Adversarial Semi-Supervised Learning Approach,
D. Kim, Jinsoo Choi, T. Oh, and I.S. Kweon,

Conference on Empirical Methods in Natural Language Processing (EMNLP) Nov. 2019.
[Paper] [Project page]

Dense Relational Captioning: Triple-Stream Networks for Relationship-Based Captioning,
D. Kim, Jinsoo Choi, T. Oh, and I.S. Kweon,

IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Jun. 2019.
[Paper] [Project page]

Contextually Customized Video Summaries via Natural Language,
Jinsoo Choi, T. Oh, and I.S. Kweon,

IEEE Winter Conference on Applications of Computer Vision (WACV), Mar. 2018.
[Paper] [Project page]

Disjoint Multi-task Learning between Heterogeneous Human-centric Tasks,
D. Kim, Jinsoo Choi, T. Oh, Y. Yoon, I.S. Kweon,

IEEE Winter Conference on Applications of Computer Vision (WACV), Mar. 2018.
[Paper]

A Real-time Vehicular Vision System to Seamlessly See-through Cars,
F. Rameau, H. Ha, K. Joo, Jinsoo Choi, I.S. Kweon,

European Conference on Computer Vision Workshops (ECCVW - ACVR), Oct. 2016.

A Real-time Augmented Reality System to See-Through Cars,
F. Rameau, H. Ha, K. Joo, Jinsoo Choi, K. Park, I.S. Kweon,

IEEE International Symposium on Mixed and Augmented Reality (ISMAR/TVCG), Sep. 2016.

Video-Story Composition via Plot Analysis,
Jinsoo Choi, T. Oh, and I.S. Kweon,

IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Jun. 2016.
Spotlight Paper (9.7% accept rate)
[Paper] [Project page] [Presentation in Las Vegas]

GMM-based Saliency Aggregation for Calibration-free Gaze Estimation,
Jinsoo Choi, B. Ahn, J. Park, and I.S. Kweon,

IEEE Conference on Image Processing (ICIP), Oct. 2014.
[Paper]

Academic Services

• CVPR Reviewer (2019, 2020)

• ICCV Reviewer (2019)

• ECCV Reviewer (2020)

• ICML Area Chair/Reviewer (2020)

• AAAI Program Committee (2020)

• TVCG Reviewer (2020)

• IEEE Access Reviewer (2019)

Awards & Honors

• Silver Prize ($7,000 grant), Samsung Humantech Paper Awards. (Feb. 2020)

• Grand Prize ($10,000 grant), The Electronic Times Paper Awards, Ministry of Science and ICT. (Nov. 2019)

• Silver Prize ($3,000 grant), Samsung Electro-Mechanics Paper Awards. (Nov. 2019)

• Qualcomm-KAIST Best Paper Award ($5,000 grant) (Nov. 2019)

• Top Research Achievements, KAIST Annual R&D Report (Nov. 2019)

• Video Intelligence Best Paper Award, AIIA Korea (Sep. 2019)

• Qualcomm-KAIST Paper Awards Nomination. (Jul. 2016)

• Grand Prize, KAIST Innovation Contest. (Jun. 2011)

• Presidential Design Award, KAIST. (Feb. 2010)