Dr. YongHo Jeong

In this study, we propose a solution for automatically measuring body circumferences by utilizing the built-in LiDAR sensor in mobile devices. Traditional body measurement methods mainly rely on 2D images or manual measurements. This research, however, utilizes 3D depth information to enhance both accuracy and efficiency. By employing HRNet-based keypoint detection and transfer learning through deep learning, the precise locations of body parts are identified and combined with depth maps to automatically calculate body circumferences. Experimental results demonstrate that the proposed method exhibits a relative error of up to 8% for major body parts such as waist, chest, hip, and buttock circumferences, with waist and buttock measurements recording low error rates below 4%. Although some models showed error rates of 7.8% and 7.4% in hip circumference measurements, this was attributed to the complexity of 3D structures and the challenges in selecting keypoint locations. Additionally, the use of depth map-based keypoint correction and regression analysis significantly improved accuracy compared to conventional 2D-based measurement methods. The real-time processing speed was also excellent, ensuring stable performance across various body types.

20 November 2023 in Physical Review D

Journal Article

First measurement of the forward rapidity gap distribution in

p Pb

collisions at

\sqrt{s_{NN}} = 8.16 TeV

CMS Collaboration

+2296

For the first time at LHC energies, the forward rapidity gap spectra from proton-lead collisions for both proton and lead dissociation processes are presented. The analysis is performed over 10.4 units of pseudorapidity at a center-of-mass energy per nucleon pair of

{\sqrt{s}}_{NN} = 8.16 TeV

, almost 300 times higher than in previous measurements of diffractive production in proton-nucleus collisions. For lead dissociation processes, which correspond to the pomeron-lead event topology, the epos-lhc generator predictions are a factor of 2 below the data, but the model gives a reasonable description of the rapidity gap spectrum shape. For the pomeron-proton topology, the epos-lhc, qgsjet ii, and hijing predictions are all at least a factor of 5 lower than the data. The latter effect might be explained by a significant contribution of ultraperipheral photoproduction events mimicking the signature of diffractive processes. These data may be of significant help in understanding the high energy limit of quantum chromodynamics and for modeling cosmic ray air showers.

8 January 2025 in Preprints.org

Preprint

Development of a 3D Point Cloud Data Analysis Model Using Mobile Devices

In this study, we propose a solution for automatically measuring body circumferences by utilizing the built-in LiDAR sensor in mobile devices. While traditional body measurement methods primarily rely on 2D images or manual measurements, this research leverages 3D depth information to enable more accurate and efficient measurements. By employing HRNet-based keypoint detection and transfer learning through deep learning, the precise locations of body parts are identified and combined with depth maps to automatically calculate body circumferences. Experimental results demonstrate that the proposed method exhibits a relative error of up to 8\% for major body parts such as waist, chest, hip, and buttock circumferences, with waist and buttock measurements recording low error rates below 4\%. Although some models showed error rates of 7.8\% and 7.4\% in hip circumference measurements, this was attributed to the complexity of 3D structures and the challenges in selecting keypoint locations. Additionally, the use of depth map-based keypoint correction and regression analysis significantly improved accuracy compared to conventional 2D-based measurement methods. The real-time processing speed was also excellent, ensuring stable performance across various body types.

14 December 2023 in Applied Sciences

Journal Article

A Multi-View Integrated Ensemble for the Background Discrimination of Semi-Supervised Semantic Segmentation

The key to semi-supervised semantic segmentation is to assign the appropriate pseudo-label to the pixels of unlabeled images. Recently, various approaches to consistency-based training and the filtering of reliable pseudo-labels have shown remarkable results. Nonetheless, there are still issues to be addressed. We find that recent approaches have specific problems in common. In pseudo-labels for training unlabeled images, we confirm that false foreground class pseudo-labels are mostly caused by background class confusion, not confusion between different foreground classes. To solve this problem, we propose a foreground and background discrimination model for semi-supervised semantic segmentation. Our proposed model is trained using a novel approach called multi-view integrated ensemble (MVIE) via output perturbation. Experimental results in various partition protocols show that our approach outperforms the existing state of the art (SOTA) in binary prediction on unlabeled data, and the segmentation model trained with the help of our model outperforms existing models.

About Me

My research journey began with my doctoral work at Sungkyunkwan University, where I focused on searching for R-parity violating supersymmetry using the CMS detector at CERN. This experience sharpened my skills in data analysis and statistical methods while fostering a deep appreciation for collaborative, international research environments. As a postdoctoral researcher at the University of Seoul and Konkuk University, I expanded my expertise into advanced AI techniques, including deep learning for 3D point cloud data analysis and computer vision applications. I am technically proficient in tools such as CMS software and ROOT for particle physics data analysis, alongside programming languages like C++ and Python. My expertise in AI and machine learning, particularly in computer vision, equips me to tackle complex challenges and develop creative solutions. Additionally, my language skills in English (IM3 level) and French (novice) enhance my ability to communicate effectively in diverse, multicultural settings. I have also contributed to the academic community through teaching and innovation. As an adjunct lecturer at Konkuk University, I taught courses on computer vision and deep learning, inspiring the next generation of scientists and engineers. Furthermore, my patent application for a dual-port interferometer underscores my commitment to advancing technology with practical applications.

Dr. YongHo Jeong

Dr. YongHo Jeong

Research Focus & Metrics

Academic Career Length

Publication Type Breakdown

Representative Publications

About Me