Search

Your search keyword '"Jang, Ji"' showing total 5,836 results
Start Over Author "Jang, Ji"
5,836 results on '"Jang, Ji"'

12. PODIA-3D: Domain Adaptation of 3D Generative Model Across Large Domain Gap Using Pose-Preserved Text-to-Image Diffusion

Author

Kim, Gwanghyun, Jang, Ji Ha, and Chun, Se Young

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Artificial Intelligence
Abstract

Recently, significant advancements have been made in 3D generative models, however training these models across diverse domains is challenging and requires an huge amount of training data and knowledge of pose distribution. Text-guided domain adaptation methods have allowed the generator to be adapted to the target domains using text prompts, thereby obviating the need for assembling numerous data. Recently, DATID-3D presents impressive quality of samples in text-guided domain, preserving diversity in text by leveraging text-to-image diffusion. However, adapting 3D generators to domains with significant domain gaps from the source domain still remains challenging due to issues in current text-to-image diffusion models as following: 1) shape-pose trade-off in diffusion-based translation, 2) pose bias, and 3) instance bias in the target domain, resulting in inferior 3D shapes, low text-image correspondence, and low intra-domain diversity in the generated samples. To address these issues, we propose a novel pipeline called PODIA-3D, which uses pose-preserved text-to-image diffusion-based domain adaptation for 3D generative models. We construct a pose-preserved text-to-image diffusion model that allows the use of extremely high-level noise for significant domain changes. We also propose specialized-to-general sampling strategies to improve the details of the generated samples. Moreover, to overcome the instance bias, we introduce a text-guided debiasing method that improves intra-domain diversity. Consequently, our method successfully adapts 3D generators across significant domain gaps. Our qualitative results and user study demonstrates that our approach outperforms existing 3D text-guided domain adaptation methods in terms of text-image correspondence, realism, diversity of rendered images, and sense of depth of 3D shapes in the generated samples, Comment: Project page: https://gwang-kim.github.io/podia_3d/

Published
2023

17. Shape-aware Text-driven Layered Video Editing

Author

Lee, Yao-Chih, Jang, Ji-Ze Genevieve, Chen, Yi-Ting, Qiu, Elizabeth, and Huang, Jia-Bin

Subjects
Computer Science - Computer Vision and Pattern Recognition, Electrical Engineering and Systems Science - Image and Video Processing
Abstract

Temporal consistency is essential for video editing applications. Existing work on layered representation of videos allows propagating edits consistently to each frame. These methods, however, can only edit object appearance rather than object shape changes due to the limitation of using a fixed UV mapping field for texture atlas. We present a shape-aware, text-driven video editing method to tackle this challenge. To handle shape changes in video editing, we first propagate the deformation field between the input and edited keyframe to all frames. We then leverage a pre-trained text-conditioned diffusion model as guidance for refining shape distortion and completing unseen regions. The experimental results demonstrate that our method can achieve shape-aware consistent video editing and compare favorably with the state-of-the-art., Comment: Project page: https://text-video-edit.github.io/

Published
2023

27. Gastrointestinal and non-gastrointestinal complication rates associated with diagnostic esophagogastroduodenoscopy under sedation.

Author

Jang, Ji, Park, Su, Yoon, Jin, Kwak, Min, and Cha, Jae

Subjects
Aged, Conscious Sedation, Endoscopy, Digestive System, Humans, Hypnotics and Sedatives, Midazolam, Myocardial Infarction, Propofol, Retrospective Studies
Abstract

Esophagogastroduodenoscopy (EGD) under sedation may result in gastrointestinal (GI) and non-GI complications. However, no previous studies have reported 30-day GI and non-GI complications after diagnostic EGD under sedation.We conducted a retrospective, observational study of 30-day GI and non-GI complication rates after outpatient diagnostic EGD under sedation in subjects ≥18 years between January 2012 and December 2017 based on a common data model database. Thirty-day complication rates were compared with EGD under sedation or not, type of sedation drugs (midazolam only vs midazolam/propofol) and age groups (18-64 year vs ≥65 year) for GI (bleeding and perforation) and non-GI complications (pneumonia, acute myocardial infarction, congestive heart failure and cerebral stroke).In total, 39,910 were performed with sedation (midazolam only, n = 16,033 and midazolam/propofol, n = 23,864) and 22,894 were performed without sedation. Elderly patients significantly favored EGD without sedation (P  .1) except for acute myocardial infarction rate, which was significantly higher in EGD without sedation than EGD under sedation (1.7/10,000 vs 0.3/10,000 persons, P = .043). All GI and non-GI complications were also similar between the midazolam/propofol and midazolam only groups as well as between young and old patients (all P > .1).Outpatient diagnostic EGD under sedation has an excellent safety profile. In addition, it can be safely performed with midazolam only or midazolam/propofol and in young and old patients.

Published
2022

37. Clinical Factors, Treatment Selection, and Outcomes in Postcataract Fungal Endophthalmitis from Contaminated Viscoelastics

Author

Nam, Dong Heun, Lee, Dae Yeong, Park, Young-Hoon, Park, Young-Gun, Kim, Mirinae, Lee, Mee Yon, Choi, Soon-Il, Kim, Min, Lee, Junwon, Moon, Sang Woong, Chung, Hyewon, Lee, Young Hoon, Lee, Min Woo, Shin, Jae Pil, Park, Dong Ho, Park, Han Sang, Kang, Yong Koo, Chung, In Young, Yoo, Woong-Sun, Cho, Yong-Wun, Kim, Yu Cheol, Kang, Kyung Tae, Jang, Ji Hye, Kim, Seong-Woo, Choi, Kwang-Eon, Choi, Mihyun, Nam, Ki Tae, Kim, Young Ho, Ahn, So Min, Lee, Sang Joon, Lim, Jae Wan, Ku, Hee Chan, Yoo, Su Jin, Choi, Moon Jung, Lew, Young Ju, Kim, Jae Hui, Kim, Joo Yeon, Hwang, Jae Hyung, Lee, Seung Woo, Bae, Kunho, Jeong, Woo Jin, Kwon, Yoon Hyung, An, Seoung Hyun, Ahn, Jeeyun, Shin, Joo Young, Byon, Ik Soo, Park, Sung Who, Park, Kyu Hyung, Kang, Se Woong, Ham, Don-Il, Kim, Sang Jin, Park, Un Chul, Oh, Baek Lok, Kim, June-Gone, Lee, Joo Yong, Lee, Jun Yeop, Kim, Yoon Jeon, Chae, Ju Byung, Kim, Dong Yoon, Lee, Christopher Seungkyu, Cho, Kwan Hyuk, Kim, Min Seok, Kim, Seong Wook, Lee, Jung Pil, Kim, Hoon Dong, Cho, In Hwan, Song, Ji Hun, Lee, Ki Hwang, Kwon, Han Jo, Lee, Seung Min, Sagong, Min, Lee, Tae Gon, Han, Jung Il, Chang, Young Suk, Min, Jung Kee, Yang, Yun Sik, Kim, Soo Han, Chin, Hee Seung, Ji, Yong Sok, Lee, Jong Young, Kim, Seong Taeck, Shin, Min Ho, Kim, Jee Taek, Han, Yong Seop, Lee, Woo Hyuk, Jo, Young Joon, Kim, Jung Yeul, Kim, Kyung Tae, Seo, Eoi Jong, Lim, Sun Taek, Lee, Jun Sung, Kim, Chang Ryong, Duck-Jin Hwang, Daniel, Kim, Ha Kyoung, Ma, Dae Joong, Hong, In Hwan, Cho, Bum Joo, Shin, Yong Un, Cho, Heeyoon, and Ahn, Seong Joon

Published
2024
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results