Zero-Shot Image Dehazing Using Pseudo Atmospheric Light Image Author: Eunsung Jo; Eunpil Park; Jae-Young Sim Abstract: Abundant training data for deep neural networks improve the performance of single image dehazing (SID) substantially, however they suffer from the domain-shift problem of the discrepancy between the training set and the...
An Empirical Study on How People Perceive AI-generated Music Authors: Hyeshin Chu, Joohee Kim, Seongouk Kim, Hongkyu Lim, Hyunwook Lee, Seungmin Jin, Jongeun Lee, Taehwan Kim, Sungahn Ko ABSTRACT Music creation is difficult because one must express one's creativity while following strict rules. The advancement of deep learning technologies...
3DMesh-GAR: 3D Human Body Mesh-Based Method for Group Activity Recognition by Muhammad Saqlain, Donguk Kim, Junuk Cha, Changhwa Lee, Seongyeong Lee and Seungryul Baek Abstract Group activity recognition is a prime research topic in video understanding and has many practical applications, such as crowd behavior monitoring, video surveillance,...
Learning 3D Skeletal Representation From Transformer for Action Recognition Junuk Cha; Muhammad Saqlain; Donguk Kim; Seungeun Lee; Seongyeong Lee; Seungryul Baek Abstract: Skeleton-based human action recognition has attracted significant interest due to its simplicity and good accuracy. Diverse end-to-end trainable frameworks based on skeletal representation have been proposed...
Centered Symmetric Quantization for Hardware-Efficient Low-Bit Neural Networks Author: Faaiz Asim, Jaewoo Park, Azat Azamat, Jongeun Lee Abstract Recent advances in quantized neural networks (QNNs) are closing the performance gap with the full precision neural networks. However at very low precision (i.e., -bits), QNNs often still suffer significant...
Authors: Azat Azamat, Jaewoo Park, Jongeun Lee ABSTRACT The cost and power consumption of BNN (Binarized Neural Network) hardware is dominated by additions. In particular, accumulators account for a large fraction of hardware overhead, which could be effectively reduced by using reduced-width accumulators. However, it is...
A Visual Analytics System for Improving Attention-based Traffic Forecasting Models Seungmin Jin, Hyunwook Lee, Cheonbok Park, Hyeshin Chu, Yunwon Tae, Jaegul Choo, Sungahn Ko With deep learning (DL) outperforming conventional methods for different tasks, much effort has been devoted to utilizing DL in various domains. Researchers and developers in the traffic domain have...
Quasi-Globally Optimal and Real-Time Visual Compass in Manhattan Structured Environments IEEE Robotics and Automation Letters ( IF 5.2 ) Pub Date: 2022-01-11 , DOI:10.1109/lra.2022.3141751 Pyojin Kim 1 , Haoang Li 2 , Kyungdon Joo 3 Affiliations We present a drift-free visual compass for estimating the three degrees of freedom (DoF) rotational motion of a camera by recognizing structural regularities in a Manhattan world...
Adaptive Cost Volume Fusion Network for Multi-Modal Depth Estimation in Changing Environments Park, Jinsun / Jeong, Yongseop / Joo, Kyungdon / Cho, Donghyeon / Kweon, In So In this letter, we propose an adaptive cost volume fusion algorithm for multi-modal depth estimation in changing environments. Our method...
Learning to Remember Patterns: Pattern Matching Memory Networks for Traffic Forecasting Hyunwook Lee, Seungmin Jin, Hyeshin Chu, Hongkyu Lim, Sungahn Ko Traffic forecasting is a challenging problem due to complex road networks and sudden speed changes caused by various events on roads. A number of models have been proposed to solve...
