Zobrazeno 1 - 10
of 142
pro vyhledávání: '"Byoung Chul Ko"'
Publikováno v:
IEEE Access, Vol 12, Pp 107385-107393 (2024)
Temporal action detection (TAD) is one of the most active research areas in computer vision. TAD is the task of detecting actions in untrimmed videos and predicting the start and end times of the actions. TAD is a challenging task and requires a vari
Externí odkaz:
https://doaj.org/article/fda71449b0cc4a549d42b52e011a2455
Publikováno v:
IEEE Access, Vol 11, Pp 59774-59787 (2023)
Facial expression recognition (FER) in the wild from various viewpoints, lighting conditions, face poses, scales, and occlusions is an extremely challenging field of research. In this study, we construct a face graph by selecting action units that pl
Externí odkaz:
https://doaj.org/article/e33a22acde77489ca34237c9727af87e
Autor:
Hyeongjin Kim, Byoung Chul Ko
Publikováno v:
Sensors, Vol 24, Iss 4, p 1111 (2024)
In this paper, we propose a new type of vision transformer (ViT) based on graph head attention (GHA). Because the multi-head attention (MHA) of a pure ViT requires multiple parameters and tends to lose the locality of an image, we replaced MHA with G
Externí odkaz:
https://doaj.org/article/e4baaa4139b5465ea80f1962999ed143
Autor:
Sangwon Kim, Byoung Chul Ko
Publikováno v:
Machine Learning: Science and Technology, Vol 5, Iss 3, p 035042 (2024)
In fields requiring high accountability, it is necessary to understand how deep-learning models make decisions when analyzing the causes of image classification. Concept-based interpretation methods have recently been introduced to reveal the interna
Externí odkaz:
https://doaj.org/article/7247f9fdf2334da3a099e84f5c62de30
Autor:
So-Hyun Cho, Su-Min Lee, Na-Young Lee, Byoung Chul Ko, Hojeong Kim, Dae-Jin Jang, Jong-Ha Lee
Publikováno v:
Sensors, Vol 23, Iss 7, p 3451 (2023)
In this study, we propose the direct diagnosis of thyroid cancer using a small probe. The probe can easily check the abnormalities of existing thyroid tissue without relying on experts, which reduces the cost of examining thyroid tissue and enables t
Externí odkaz:
https://doaj.org/article/1d534749d51a4e48b1e591547a1321da
Publikováno v:
IEEE Access, Vol 9, Pp 40850-40859 (2021)
This paper proposes a keypoint regressor (KeyReg), which consists of multi-layer random forest (MRF) regressor and single random forest (SRF) classifier modules. To increase the keypoints’ repeatability, the MRF regressor is applied to multi-scale
Externí odkaz:
https://doaj.org/article/114423a5af5141a8b96bce400a5993eb
Publikováno v:
IEEE Access, Vol 9, Pp 114535-114546 (2021)
In this paper, a graph convolutional network (GCN)-based multi-object tracking (MOT) algorithm, consisting of a module for extracting the initial features and a module for updating the features, that estimates the affinity between nodes is proposed.
Externí odkaz:
https://doaj.org/article/b1688d0bb9e0427e830ad3830a80c3e9
Autor:
Sangwon Kim, Byoung Chul Ko
Publikováno v:
IEEE Access, Vol 8, Pp 8533-8542 (2020)
Randomized sampling-based ensemble learning is emerging as a new alternative to deep neural networks (DNNs) because it supports diversity and locality and does not require backpropagation in the learning process. By connecting randomized weak classif
Externí odkaz:
https://doaj.org/article/555d2f99acc2496396ed7b4394ca89ee
Publikováno v:
IEEE Access, Vol 8, Pp 60344-60354 (2020)
This study proposes a lightweight multilayer random forest (LMRF) model, which is a non-neural network style deep model consisting of layer-by-layer random forests. Although a deep neural network (DNN) is a powerful algorithm for facial expression re
Externí odkaz:
https://doaj.org/article/70382c404fd940c4ba62164fa0263655
Publikováno v:
IEEE Access, Vol 8, Pp 182828-182841 (2020)
In a multiple object tracking (MOT) system, an association check between the tracker and detected objects is an important factor in determining the tracking performance. Siamese convolution neural network (CNN) is the most popular data association me
Externí odkaz:
https://doaj.org/article/f0762a43a630464f81b7cfc1757a30ce