Junsik Kim

Junsik Kim

KAIST

H-index: 15

Asia-South Korea

About Junsik Kim

Junsik Kim, With an exceptional h-index of 15 and a recent h-index of 15 (since 2020), a distinguished researcher at KAIST, specializes in the field of Computer Vision, Machine Learning.

His recent articles reflect a diverse array of research interests and contributions to the field:

Joint-Task Regularization for Partially Labeled Multi-Task Learning

Blurry Video Compression: A Trade-Off Between Visual Enhancement and Data Compression

ENInst: Enhancing weakly-supervised low-shot instance segmentation

Weakly supervised contrastive learning for unsupervised vehicle reidentification

An Iterative Method for Unsupervised Robust Anomaly Detection Under Data Contamination

Active anomaly detection based on deep one-class classification

Sound source localization is all about cross-modal alignment

Event-specific audio-visual fusion layers: A simple and new perspective on video understanding

Junsik Kim Information

University

Position

___

Citations(all)

1519

Citations(since 2020)

1448

Cited By

562

hIndex(all)

15

hIndex(since 2020)

15

i10Index(all)

18

i10Index(since 2020)

18

Email

University Profile Page

Google Scholar

Junsik Kim Skills & Research Interests

Computer Vision

Machine Learning

Top articles of Junsik Kim

Joint-Task Regularization for Partially Labeled Multi-Task Learning

arXiv preprint arXiv:2404.01976

2024/4/2

Blurry Video Compression: A Trade-Off Between Visual Enhancement and Data Compression

2024

Junsik Kim
Junsik Kim

H-Index: 9

In So Kweon
In So Kweon

H-Index: 53

ENInst: Enhancing weakly-supervised low-shot instance segmentation

Pattern Recognition

2024/1/1

Junsik Kim
Junsik Kim

H-Index: 9

Weakly supervised contrastive learning for unsupervised vehicle reidentification

IEEE Transactions on Image Processing

2024/3/27

An Iterative Method for Unsupervised Robust Anomaly Detection Under Data Contamination

IEEE Transactions on Neural Networks and Learning Systems

2023/5/9

Active anomaly detection based on deep one-class classification

Pattern Recognition Letters

2023/3/1

Sound source localization is all about cross-modal alignment

2023

Event-specific audio-visual fusion layers: A simple and new perspective on video understanding

2023

Ml-bpm: Multi-teacher learning with bidirectional photometric mixing for open compound domain adaptation in semantic segmentation

2022/10/22

Labeling Where Adapting Fails: Cross-Domain Semantic Segmentation with Point Supervision via Active Selection

arXiv preprint arXiv:2206.00181

2022/6/1

Learning sound localization better from semantically similar samples

2022/5/23

Camera-tracklet-aware contrastive learning for unsupervised vehicle re-identification

2022/5/23

How does AI improve human decision-making? Evidence from the AI-powered Go program

Evidence from the AI-Powered Go Program (April 2022). USC Marshall School of Business Research Paper Sponsored by iORB, No. Forthcoming

2022/4/1

Audio-visual fusion layers for event type aware video recognition

arXiv preprint arXiv:2202.05961

2022/2/12

Less can be more: Sound source localization with a classification model

2022

Unsupervised deep one-class classification with adaptive threshold based on training dynamics

2022/11/28

Normality-Calibrated Autoencoder for Unsupervised Anomaly Detection on Data Contamination

NeurIPS 2021 Workshop DGMs Applications

2021/10/14

Motion-blurred video interpolation and extrapolation

The Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI-21)

2021/3/4

Junsik Kim
Junsik Kim

H-Index: 9

In So Kweon
In So Kweon

H-Index: 53

Optical flow estimation from a single motion-blurred image

The Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI-21)

2021/3/4

Restoration of video frames from a single blurred image with motion understanding

2021

See List of Professors in Junsik Kim University(KAIST)

Co-Authors

academic-engine