Filters








98 Hits in 1.5 sec

DTM: Deformable Template Matching [article]

Hyungtae Lee, Heesung Kwon, Ryan M. Robinson, William D. Nothwang
2016 arXiv   pre-print
A novel template matching algorithm that can incorporate the concept of deformable parts, is presented in this paper. Unlike the deformable part model (DPM) employed in object recognition, the proposed template-matching approach called Deformable Template Matching (DTM) does not require a training step. Instead, deformation is achieved by a set of predefined basic rules (e.g. the left sub-patch cannot pass across the right patch). Experimental evaluation of this new method using the PASCAL VOC
more » ... 7 dataset demonstrated substantial performance improvement over conventional template matching algorithms. Additionally, to confirm the applicability of DTM, the concept is applied to the generation of a rotation-invariant SIFT descriptor. Experimental evaluation employing deformable matching of SIFT features shows an increased number of matching features compared to a conventional SIFT matching.
arXiv:1604.03518v1 fatcat:uovuiodu7ranniykbifquhzc3i

Is Pretraining Necessary for Hyperspectral Image Classification? [article]

Hyungtae Lee, Sungmin Eum, Heesung Kwon
2019 arXiv   pre-print
Lee et al. [2] introduced Cross-Domain CNN which simultaneously trains multiple networks for classifying multiple hyperspectral image domains.  ... 
arXiv:1901.08658v1 fatcat:2wequy42kbfqvajd4zeairiuia

Self-supervised Contrastive Learning for Cross-domain Hyperspectral Image Representation [article]

Hyungtae Lee, Heesung Kwon
2022 arXiv   pre-print
Recently, self-supervised learning has attracted attention due to its remarkable ability to acquire meaningful representations for classification tasks without using semantic labels. This paper introduces a self-supervised learning framework suitable for hyperspectral images that are inherently challenging to annotate. The proposed framework architecture leverages cross-domain CNN, allowing for learning representations from different hyperspectral images with varying spectral characteristics
more » ... no pixel-level annotation. In the framework, cross-domain representations are learned via contrastive learning where neighboring spectral vectors in the same image are clustered together in a common representation space encompassing multiple hyperspectral images. In contrast, spectral vectors in different hyperspectral images are separated into distinct clusters in the space. To verify that the learned representation through contrastive learning is effectively transferred into a downstream task, we perform a classification task on hyperspectral images. The experimental results demonstrate the advantage of the proposed self-supervised representation over models trained from scratch or other transfer learning methods.
arXiv:2202.03968v1 fatcat:k22cxyndi5dxdlnvdgwba4q4yu

Dynamic Belief Fusion for Object Detection [article]

Hyungtae Lee, Heesung Kwon, Ryan M. Robinson, William d. Nothwang, and Amar M. Marathe
2015 arXiv   pre-print
Kwon and Lee proposed two approaches integrating multiple sample-based tracking approaches using an interactive Markov Chain Monte Carlo (iMCMC) framework [12] and using sampling in tracker space modeled  ... 
arXiv:1511.03183v1 fatcat:5umpydybpnczlldrvzpzbozhva

IOD-CNN: Integrating Object Detection Networks for Event Recognition [article]

Sungmin Eum, Hyungtae Lee, Heesung Kwon, David Doermann
2017 arXiv   pre-print
Many previous methods have showed the importance of considering semantically relevant objects for performing event recognition, yet none of the methods have exploited the power of deep convolutional neural networks to directly integrate relevant object information into a unified network. We present a novel unified deep CNN architecture which integrates architecturally different, yet semantically-related object detection networks to enhance the performance of the event recognition task. Our
more » ... tecture allows the sharing of the convolutional layers and a fully connected layer which effectively integrates event recognition, rigid object detection and non-rigid object detection.
arXiv:1703.07431v1 fatcat:h5hjy67gbjdv3p2bq5vwjza3da

Cross-domain CNN for Hyperspectral Image Classification [article]

Hyungtae Lee, Sungmin Eum, Heesung Kwon
2018 arXiv   pre-print
For each dataset-specific stream in the proposed architecture, we have used the modified version of the 9-layered hyperspectral image classification CNN introduced by Lee and Kwon [4, 5] .  ... 
arXiv:1802.00093v2 fatcat:f7wc6qz3p5hnzczl4u52p6qade

Enhanced Object Detection via Fusion With Prior Beliefs from Image Classification [article]

Yilun Cao and Hyungtae Lee and Heesung Kwon
2016 arXiv   pre-print
Lee et al.  ...  Dynamic Belief Fusion (DBF) To effectively fuse detection scores of each detection window from individual detector and classifier, a novel fusion method proposed by Lee et al.  ... 
arXiv:1610.06907v1 fatcat:pubi5cg5evboddyr27sa52t6da

Cervical plexus block

Jin-Soo Kim, Justin Sangwook Ko, Seunguk Bang, Hyungtae Kim, Sook Young Lee
2018 Korean Journal of Anesthesiology  
-0003-2488-9986 Sook Young Lee, https://orcid.org/0000-0002-4688-2155  ...  ORCID Jin-Soo Kim, https://orcid.org/0000-0003-4121-2475 Justin Sangwook Ko, https://orcid.org/0000-0003-3155-0550 Seunguk Bang, https://orcid.org/0000-0001-6609-7691 Hyungtae Kim, https://orcid.org/0000  ...  Intermediate cervical plexus block History In 2002, Zhang and Lee [21] reported that the investing layer of the deep fascia does not exist in the space between the SCM and trapezius muscles, an area  ... 
doi:10.4097/kja.d.18.00143 pmid:29969890 pmcid:PMC6078883 fatcat:hddyyv5xgvaddd7hqrvb4klpgi

Effects of Female Reproductive Hormone Levels on Inadvertent Intraoperative Hypothermia during Laparoscopic Gynecologic Surgery: A Retrospective Study

Cheol Lee, SeongNam Park, ByoungRyun Kim, Hyeonbin Yim, Myeongjong Lee, Juhwan Lee, Hyungtae Kim
2021 Medicina  
and Objectives: Female reproductive hormones may affect core body temperature. This study aimed to investigate the effects of female reproductive hormones on inadvertent intraoperative hypothermia in patients who underwent laparoscopic gynecologic surgery under general anesthesia. Materials and Methods: This retrospective study included 660 menstruating and menopausal female patients aged 19–65 years. The patients were divided into two groups according to the occurrence of inadvertent
more » ... tive hypothermia: non-hypothermia group (N = 472) and hypothermia group (N = 188). After propensity score matching, 312 patients (N = 156 in each group) were analyzed to investigate the association between intraoperative hypothermia and female reproductive hormones. As potential predictors of inadvertent hypothermia, the levels of female reproductive hormones were analyzed using binary logistic regression. Results: The association of estradiol (r = −0.218, p = 0.000) and progesterone (r = −0.235, p = 0.000) levels with inadvertent intraoperative hypothermia was significant but weakly negative before matching; however, it was significant and moderately negative after matching (r = −0.326, p = 0.000 and r = −0.485, p = 0.000, respectively). In a binary logistic analysis, the odds ratio for estradiol was 0.995 (p = 0.014, 0.993 < 95% confidence interval [CI] < 0.998) before matching and 0.993 (p = 0.000, 0.862 < 95% CI < 0.930) after matching, and that for progesterone was 0.895 (p = 0.000, 0.862 < 95% CI < 0.930) before matching and 0.833 (p = 0.014, 0.990 < 95% CI < 0.996) after matching. Conclusions: Estradiol and progesterone levels were associated with inadvertent intraoperative hypothermia. However, the odds ratio for female reproductive hormone levels was close to 1. Therefore, female reproductive hormones may not be a risk factor for hypothermia during gynecologic surgery under general anesthesia. However, a small sample size in this study limits the generalizability of the results.
doi:10.3390/medicina57111255 pmid:34833473 pmcid:PMC8622736 fatcat:6kv4zuar45fu7jueltxgdbzgpu

Fast Object Localization Using a CNN Feature Map Based Multi-Scale Search [article]

Hyungtae Lee, Heesung Kwon, Archith J. Bency, William D. Nothwang
2016 arXiv   pre-print
Object localization is an important task in computer vision but requires a large amount of computational power due mainly to an exhaustive multiscale search on the input image. In this paper, we describe a near real-time multiscale search on a deep CNN feature map that does not use region proposals. The proposed approach effectively exploits local semantic information preserved in the feature map of the outermost convolutional layer. A multi-scale search is performed on the feature map by
more » ... sing all the sub-regions of different sizes using separate expert units of fully connected layers. Each expert unit receives as input local semantic features only from the corresponding sub-regions of a specific geometric shape. Therefore, it contains more nearly optimal parameters tailored to the corresponding shape. This multi-scale and multi-aspect ratio scanning strategy can effectively localize a potential object of an arbitrary size. The proposed approach is fast and able to localize objects of interest with a frame rate of 4 fps while providing improved detection performance over the state-of-the art on the PASCAL VOC 12 and MSCOCO data sets.
arXiv:1604.03517v1 fatcat:7eubeju3sbhc7oyf2oc4cbuwni

Weakly Supervised Localization using Deep Feature Maps [article]

Archith J. Bency, Heesung Kwon, Hyungtae Lee, S. Karthikeyan, B. S. Manjunath
2016 arXiv   pre-print
Object localization is an important computer vision problem with a variety of applications. The lack of large scale object-level annotations and the relative abundance of image-level labels makes a compelling case for weak supervision in the object localization task. Deep Convolutional Neural Networks are a class of state-of-the-art methods for the related problem of object recognition. In this paper, we describe a novel object localization algorithm which uses classification networks trained
more » ... only image labels. This weakly supervised method leverages local spatial and semantic patterns captured in the convolutional layers of classification networks. We propose an efficient beam search based approach to detect and localize multiple objects in images. The proposed method significantly outperforms the state-of-the-art in standard object localization data-sets with a 8 point increase in mAP scores.
arXiv:1603.00489v2 fatcat:s2yajbnnvjaibfy4kfcqfmmif4

DOD-CNN: Doubly-injecting Object Information for Event Recognition [article]

Hyungtae Lee, Sungmin Eum, Heesung Kwon
2019 arXiv   pre-print
Recognizing an event in an image can be enhanced by detecting relevant objects in two ways: 1) indirectly utilizing object detection information within the unified architecture or 2) directly making use of the object detection output results. We introduce a novel approach, referred to as Doubly-injected Object Detection CNN (DOD-CNN), exploiting the object information in both ways for the task of event recognition. The structure of this network is inspired by the Integrated Object Detection CNN
more » ... (IOD-CNN) where object information is indirectly exploited by the event recognition module through the shared portion of the network. In the DOD-CNN architecture, the intermediate object detection outputs are directly injected into the event recognition network while keeping the indirect sharing structure inherited from the IOD-CNN, thus being 'doubly-injected'. We also introduce a batch pooling layer which constructs one representative feature map from multiple object hypotheses. We have demonstrated the effectiveness of injecting the object detection information in two different ways in the task of malicious event recognition.
arXiv:1811.02910v2 fatcat:cto6nzicqja5xpjvxt467lhuze

A Single Correspondence Is Enough: Robust Global Registration to Avoid Degeneracy in Urban Environments [article]

Hyungtae Lim, Suyong Yeon, Soohyun Ryu, Yonghan Lee, Youngji Kim, Jaeseong Yun, Euigon Jung, Donghwan Lee, Hyun Myung
2022 arXiv   pre-print
Global registration using 3D point clouds is a crucial technology for mobile platforms to achieve localization or manage loop-closing situations. In recent years, numerous researchers have proposed global registration methods to address a large number of outlier correspondences. Unfortunately, the degeneracy problem, which represents the phenomenon in which the number of estimated inliers becomes lower than three, is still potentially inevitable. To tackle the problem, a degeneracy-robust
more » ... ling-based global registration method is proposed, called Quatro. In particular, our method employs quasi-SO(3) estimation by leveraging the Atlanta world assumption in urban environments to avoid degeneracy in rotation estimation. Thus, the minimum degree of freedom (DoF) of our method is reduced from three to one. As verified in indoor and outdoor 3D LiDAR datasets, our proposed method yields robust global registration performance compared with other global registration methods, even for distant point cloud pairs. Furthermore, the experimental results confirm the applicability of our method as a coarse alignment. Our code is available: https://github.com/url-kaist/quatro.
arXiv:2203.06612v1 fatcat:mtqgkiwla5ephhd72a4eqv3j7m

Qualitative Pose Estimation by Discriminative Deformable Part Models [chapter]

Hyungtae Lee, Vlad I. Morariu, Larry S. Davis
2013 Lecture Notes in Computer Science  
We present a discriminative deformable part model for the recovery of qualitative pose, inferring coarse pose labels (e.g., left, frontright, back), a task which we expect to be more robust to common confounding factors that hinder the inference of exact 2D or 3D joint locations. Our approach automatically selects parts that are predictive of qualitative pose and trains their appearance and deformation costs to best discriminate between qualitative poses. Unlike previous approaches, our parts
more » ... e both selected and trained to improve qualitative pose discrimination and are shared by all the qualitative pose models. This leads to both increased accuracy and higher efficiency, since fewer parts models are evaluated for each image. In comparisons with two state-of-the-art approaches on a public dataset, our model shows superior performance.
doi:10.1007/978-3-642-37444-9_57 fatcat:vut55lbhgbfkfjtlpt4sjc2bi4

JH2R: Joint Homography Estimation for Highlight Removal

Sungmin Eum, Hyungtae Lee, David Doermann
2015 Procedings of the British Machine Vision Conference 2015  
doi:10.5244/c.29.49 dblp:conf/bmvc/EumLD15 fatcat:upb5tghdqncvpksgsp3bfjvrty
« Previous Showing results 1 — 15 out of 98 results