no code implementations • 24 Apr 2024 • Chuong Huynh, Seoung Wug Oh, Abhinav Shrivastava, Joon-Young Lee
Human matting is a foundation task in image and video processing, where human foreground pixels are extracted from the input.
1 code implementation • 8 Dec 2023 • Hanjung Kim, Jaehyun Kang, Miran Heo, Sukjun Hwang, Seoung Wug Oh, Seon Joo Kim
By effectively resolving the over-reliance on location information, we achieve state-of-the-art results on YouTube-VIS 2019/2021 and Occluded VIS (OVIS).
1 code implementation • 19 Oct 2023 • Ho Kei Cheng, Seoung Wug Oh, Brian Price, Joon-Young Lee, Alexander Schwing
We present Cutie, a video object segmentation (VOS) network with object-level memory reading, which puts the object representation from memory back into the video object segmentation result.
Ranked #1 on Semi-Supervised Video Object Segmentation on MOSE
1 code implementation • ICCV 2023 • Ho Kei Cheng, Seoung Wug Oh, Brian Price, Alexander Schwing, Joon-Young Lee
To 'track anything' without training on video data for every individual task, we develop a decoupled video segmentation approach (DEVA), composed of task-specific image-level segmentation and class/task-agnostic bi-directional temporal propagation.
Ranked #1 on Unsupervised Video Object Segmentation on DAVIS 2016 val (using extra training data)
Open-Vocabulary Video Segmentation Open-World Video Segmentation +7
no code implementations • 9 Feb 2023 • Yiran Xu, Zhixin Shu, Cameron Smith, Seoung Wug Oh, Jia-Bin Huang
3D-aware GANs offer new capabilities for view synthesis while preserving the editing functionalities of their 2D counterparts.
no code implementations • CVPR 2023 • KwanYong Park, Sanghyun Woo, Seoung Wug Oh, In So Kweon, Joon-Young Lee
Mask-guided matting has shown great practicality compared to traditional trimap-based methods.
no code implementations • 20 Dec 2022 • Sanghyun Woo, KwanYong Park, Seoung Wug Oh, In So Kweon, Joon-Young Lee
First, no tracking supervisions are in LVIS, which leads to inconsistent learning of detection (with LVIS and TAO) and tracking (only with TAO).
no code implementations • 20 Dec 2022 • Sanghyun Woo, KwanYong Park, Seoung Wug Oh, In So Kweon, Joon-Young Lee
The tracking-by-detection paradigm today has become the dominant method for multi-object tracking and works by detecting objects in each frame and then performing data association across frames.
1 code implementation • CVPR 2023 • Miran Heo, Sukjun Hwang, Jeongseok Hyun, Hanjung Kim, Seoung Wug Oh, Joon-Young Lee, Seon Joo Kim
Notably, we greatly outperform the state-of-the-art on the long VIS benchmark (OVIS), improving 5. 6 AP with ResNet-50 backbone.
Ranked #6 on Video Instance Segmentation on YouTube-VIS 2021 (using extra training data)
1 code implementation • CVPR 2022 • KwanYong Park, Sanghyun Woo, Seoung Wug Oh, In So Kweon, Joon-Young Lee
In this per-clip inference scheme, we update the memory with an interval and simultaneously process a set of consecutive frames (i. e. clip) between the memory updates.
1 code implementation • 27 Jul 2022 • Hongje Seong, Seoung Wug Oh, Brian Price, Euntai Kim, Joon-Young Lee
A key of OTVM is the joint modeling of trimap propagation and alpha prediction.
no code implementations • 21 Jul 2022 • Jaeyeon Kang, Seoung Wug Oh, Seon Joo Kim
The key to video inpainting is to use correlation information from as many reference frames as possible.
1 code implementation • 9 Jun 2022 • Miran Heo, Sukjun Hwang, Seoung Wug Oh, Joon-Young Lee, Seon Joo Kim
Specifically, we use an image object detector as a means of distilling object-specific contexts into object tokens.
Ranked #11 on Video Instance Segmentation on YouTube-VIS 2021 (using extra training data)
1 code implementation • CVPR 2022 • Sukjun Hwang, Miran Heo, Seoung Wug Oh, Seon Joo Kim
The set classifier is plug-and-playable to existing object trackers, and highly improves the performance of long-tailed object tracking.
1 code implementation • CVPR 2022 • Su Ho Han, Sukjun Hwang, Seoung Wug Oh, Yeonchool Park, Hyunwoo Kim, Min-Jung Kim, Seon Joo Kim
We also introduce cooperatively operating modules that aggregate information from available frames, in order to enrich the features for all subtasks in VIS.
1 code implementation • ICCV 2021 • Hongje Seong, Seoung Wug Oh, Joon-Young Lee, Seongwon Lee, Suhyeon Lee, Euntai Kim
Based on a recent memory-based method [33], we propose two advanced memory read modules that enable us to perform memory reading in multiple scales while exploiting temporal smoothness.
1 code implementation • CVPR 2021 • Younghyun Jo, Seoung Wug Oh, Peter Vajda, Seon Joo Kim
By the one-to-many nature of the super-resolution (SR) problem, a single low-resolution (LR) image can be mapped to many high-resolution (HR) images.
Ranked #3 on Blind Super-Resolution on DIV2KRK - 4x upscaling
1 code implementation • NeurIPS 2021 • Sukjun Hwang, Miran Heo, Seoung Wug Oh, Seon Joo Kim
We propose a novel end-to-end solution for video instance segmentation (VIS) based on transformers.
Ranked #32 on Video Instance Segmentation on YouTube-VIS validation
1 code implementation • CVPR 2021 • Gunhee Nam, Miran Heo, Seoung Wug Oh, Joon-Young Lee, Seon Joo Kim
Since the existing datasets are not suitable to validate our method, we build a new polygonal point set tracking dataset and demonstrate the superior performance of our method over the baselines and existing contour-based VOS methods.
3 code implementations • CVPR 2021 • Jaedong Hwang, Seoung Wug Oh, Joon-Young Lee, Bohyung Han
We extend panoptic segmentation to the open-world and introduce an open-set panoptic segmentation (OPS) task.
no code implementations • 3 Dec 2020 • Sukjun Hwang, Seoung Wug Oh, Seon Joo Kim
Panoptic segmentation, which is a novel task of unifying instance segmentation and semantic segmentation, has attracted a lot of attention lately.
no code implementations • ECCV 2020 • Subin Jeon, Seonghyeon Nam, Seoung Wug Oh, Seon Joo Kim
To reduce the training-testing discrepancy of the self-supervised learning, a novel cross-identity training scheme is additionally introduced.
1 code implementation • ECCV 2020 • Jaeyeon Kang, Younghyun Jo, Seoung Wug Oh, Peter Vajda, Seon Joo Kim
Video super-resolution (VSR) and frame interpolation (FI) are traditional computer vision problems, and the performance have been improving by incorporating deep learning recently.
no code implementations • 20 Mar 2020 • Younghyun Jo, Jaeyeon Kang, Seoung Wug Oh, Seonghyeon Nam, Peter Vajda, Seon Joo Kim
Our framework is similar to GANs in that we iteratively train two networks - a generator and a loss network.
no code implementations • 20 Mar 2020 • Gunhee Nam, Seoung Wug Oh, Joon-Young Lee, Seon Joo Kim
We propose a novel memory-based tracker via part-level dense memory and voting-based retrieval, called DMV.
1 code implementation • ICCV 2019 • Sungho Lee, Seoung Wug Oh, DaeYeun Won, Seon Joo Kim
We propose a novel DNN-based framework called the Copy-and-Paste Networks for video inpainting that takes advantage of additional information in other frames of the video.
Ranked #6 on Video Inpainting on DAVIS
1 code implementation • ICCV 2019 • Seoung Wug Oh, Sungho Lee, Joon-Young Lee, Seon Joo Kim
Given a set of reference images and a target image with holes, our network fills the hole by referring the contents in the reference images.
1 code implementation • CVPR 2019 • Seoung Wug Oh, Joon-Young Lee, Ning Xu, Seon Joo Kim
We propose a new multi-round training scheme for the interactive video object segmentation so that the networks can learn how to understand the user's intention and update incorrect estimations during the training.
Ranked #6 on Interactive Video Object Segmentation on DAVIS 2017 (AUC-J metric)
3 code implementations • ICCV 2019 • Seoung Wug Oh, Joon-Young Lee, Ning Xu, Seon Joo Kim
In our framework, the past frames with object masks form an external memory, and the current frame as the query is segmented using the mask information in the memory.
Ranked #4 on Interactive Video Object Segmentation on DAVIS 2017 (using extra training data)
1 code implementation • CVPR 2018 • Younghyun Jo, Seoung Wug Oh, Jaeyeon Kang, Seon Joo Kim
We propose a novel end-to-end deep neural network that generates dynamic upsampling filters and a residual image, which are computed depending on the local spatio-temporal neighborhood of each pixel to avoid explicit motion compensation.
Ranked #6 on Video Super-Resolution on Vid4 - 4x upscaling
2 code implementations • CVPR 2018 • Seoung Wug Oh, Joon-Young Lee, Kalyan Sunkavalli, Seon Joo Kim
We validate our method on four benchmark sets that cover single and multiple object segmentation.
no code implementations • 29 Aug 2016 • Seoung Wug Oh, Seon Joo Kim
Computational color constancy refers to the problem of computing the illuminant color so that the images of a scene under varying illumination can be normalized to an image under the canonical illumination.
no code implementations • CVPR 2016 • Seoung Wug Oh, Michael S. Brown, Marc Pollefeys, Seon Joo Kim
In particular, due to the differences in spectral sensitivities of the cameras, different cameras yield different RGB measurements for the same spectral signal.