Sungwoo Park

Dokyeong Kwon

Suhyeon Ha

Jinhee Park

Juhyeon Park

Jin Hong

Youjin Kim

Jinwook Paeng

Guisik Kim

Dong Wook Shu

Sungmin Cho

Seohyeon Lee

Mijoo Kim

Jisoo Park

Seonghak Lee

Taeeun Kwon

Minho Lee

Dagyeong Na

Chan-Ju Park

Rae Hyuk Jung

Minseok Song

Ji Young Park

Ga Young Kim

Bo Won Choi

Hjunjun Joo

Shinyeong Yun

Jeongmin Lee

HyukJin Jang

Younkyoung Yoon

Heebin Yoo

Ji Soo Kim

Hyeseong Kim

Hyomin Kim

Wonbeom Jang

Hongchan Yoon

Kevin Jin

Minjun Kim

Dongha Lee

The Computer Vision Machine Learning Lab (CVML@CAU) at Chung-Ang University conducts research in machine learning, computer vision, and speech processing, covering topics such as confidence calibration, out-of-distribution detection, metric learning, geometric learning, adversarial attacks, representation learning, Monte Carlo sampling, variational inference, generative learning, graph learning, meta-learning, diffusion models, NeRF, 3D talking head synthesis, humanRF, object detection, 3D point clouds, event summarization, dehazing, low-light enhancement, speech separation, speech enhancement, and speech recognition. Additionally, our laboratory aims to solve problems in computer vision, object recognition, detection, and tracking using machine learning techniques. We primarily apply deep learning and big data machine learning methods to image data to address these challenges. Ultimately, computer vision and machine learning techniques developed here will be applied in diverse fields such as autonomous navigation, virtual reality, and medical imaging.

우리 연구실은 머신러닝 기술을 이용하여 컴퓨터 비전, 객체 인식, 탐지 및 추적 문제를 해결하는 것을 목표로 합니다. 최근 활발히 연구되고 있는 딥러닝 기술과 빅데이터를 활용한 머신러닝 기법을 주로 이미지 데이터에 적용하여 주어진 문제를 해결할 것입니다. 궁극적으로 컴퓨터 비전 및 머신러닝 기술은 자율 주행, 가상 현실, 의료 영상과 같은 다양한 분야에 활용될 것입니다.

Calibration-Free Multi-Object Tracking via Scale-Invariant Topological Depth and Adaptive Uncertainty Gating

WildTalker ∞ : Pushing the Limits of 3D Talking Portrait Synthesis in Unconstrained Environments

Spectral-guided Physical Dynamics Distillation

DiffMusic: Efficient Music Generation From a Single Image Using Diffusion-Based Representations

Neural Collapse-Informed Initialization with Perturbation Injection in Classification-based Metric Learning

Towards Generalizable Time Series Forecasting via IB-Regularized Transformer-Diffusion

Lightweight Wasserstein Audio-Visual Model for Unified Speech Enhancement and Separation

Occlusion-Robust Multi-Object Tracking with Adaptive Feature Management and Motion Compensation

Physics-Informed Neural Operators for Tissue Elasticity Reconstruction

Deep Disentangled Metric Learning

Text-Centric Multimodal Alignment via Dual-Level Optimization

Orthogonal Disentanglement for Generalizable Audio Deepfake Detection

Lightweight Grounding Model Combining a CLIP-based Encoder With an Upsampling Decoder

diffMusic: A Multi-Stage Model for Contextually Aligned Music Generation from Visual Inputs

MRDiff: Time Series Anomaly Detection Using Multi-level Reconstruction Diffusion

Optimizing Dataset Distillation Using DATM: Adjusting Learning Rate and Upper Bound

Satellite Image Dehazing Via Masked Image Modeling and Jigsaw Transformation

WildTalker: Talking Portrait Synthesis In the Wild

Geometry-Preserved Image Editing

POTF: Prompt-based Object-centric Tensorial Field

Uncertainty Calibration With Energy Based Instance-wise Scaling In the Wild Dataset

LOAD: Local Attribute Debiasing for Pretrained Vision Models

Deep Dehazing Powered by Image Processing Network

Riemannian Neural SDE: Learning Stochastic Representations on Manifolds

Learning to Intrinsic Image Filter for Instagram Filter Removal

B-Cell Linear Epitope Prediction Using Transformer Encoder

Balanced Data Augmentation of Object Detection Via Boot-strapping

Neural Markov Controlled SDE: Stochastic Optimization for Continuous-Time Data

Densely-packed Object Detection via Hard Negative-Aware Anchor Attention

Automatic Stylized Plaid Check Pattern Try-on

Filter Pruning via Softmax Attention

Wasserstein Distributional Normalization For Robust Distributional Certification of Noisy Labeled Data

Generative Adversarial Networks for Markovian Temporal Dynamics: Stochastic Continuous Data Generation

Dog Noseprint Identification Algorithm

Deep Diffusion-Invariant Wasserstein Distributional Classification

Visual Tracking by TridentAlign and Context Embedding

DALE : Dark Region-Aware Low-light Image Enhancement

Upright Adjustment with Graph Convolutional Networks

3D Point Cloud Generative Adversarial Network Based on Tree Structured Graph Convolutions

Deep Meta Learning for Real-Time Target-Aware Visual Tracking

Bidirectional Deep Residual learning for Haze Removal

Multi-Domain Attentive Detection Network

LOW-LIGHTGAN: Low-Light Enhancement via Advanced Generative Adversarial Network with Task-Driven Training

Learning to Remember Past to Predict Future for Visual Tracking

Depth-Controllable Very Deep Super-Resolution Network

Sphere Generative Adversarial Network Based on Geometric Moment Matching

Low-Complexity Online Model Selection with Lyapunov Control for Reward Reward Maximization in Stabilized Real-Time Deep Learning Platforms

Adaptive Patch based Convolutional Neural Network for Robust Dehazing

Robust Pixel-wise Dehazing Algorithm based on Advanced Haze-Relevant Features

Appearances can be deceiving: Learning visual tracking from few trajectory annotations

Robust Visual Tracking with Double Bounding Box Model

Interval Tracker: Tracking by Interval Analysis

Visual Tracking using Joint Inference of Target State and Segment-based Appearance Models

Minimum Uncertainty Gap for Robust Visual Tracking

Robust Visual Tracking using Autoregressive Hidden Markov Model

A Unified Framework for Event Summarization and Rare Event Detection

Tracking by Sampling Trackers

Visual Tracking Decomposition

Simultaneous Video Synchronization and Rare Event Detection via Cross-EntropyMonte Carlo Optimization

Tracking of a Non-Rigid Object via Patch-based Dynamic Appearance Modeling and Adaptive Basin Hopping Monte Carlo Sampling

Tracking of Abrupt Motion using Wang-Landau Monte Carlo Estimation

Confidence Controls Deep Metric Learning

The Competition of Fairness in AI-generated Face Detection: Methods and Result

Rethinking Metric Learning: Enhancing Generalization to Unseen Classes

VT-Surf: Visual Tracking with Switching Dynamics under Time-Series Forecasting

Enhancing Global and Local Context Modeling in Time Series Through Multi-Step Transformer-Diffusion Interaction

SOTA: Sequential Optimal Transport Approximation For Visual Tracking in Wild Scenario

Hierarchical Bidirected Graph Convolutions For Large-Scale 3D Point Cloud Place Recognition

Satellite Image Dehazing based on Dual Frequency Pass Networks

End-to-End Metric Learning From Corrupted Images Using Triplet Dimensionality Reduction Loss

Sampling Based Spherical Transformer For 360 Degree Image Classification

AttSec: Protein Secondary Structure Prediction By Capturing Local Patterns From Attention Map

Localized Curvature-based Combinatorial Subgraph Sampling for Large-scale Graphs

Dual Gradient based Snow Attentive Desnowing

Self-Parameter Distillation Dehazing

Self-Augmentation Based on Noise-Robust Probabilistic Model For Noisy Labels

Wasserstein Distributional Harvesting for Highly Dense 3D Point Clouds

Optimal Visual Tracking Using Wasserstein Transport Proposals

Wasserstein Approximate Bayesian Computation for Visual Tracking

Style Transfer with Target Feature Palette and Attention Coloring

Stagemix video generation using face and body keypoints detection

Adversarial Attack Can Help Visual Tracking

Riemannian Submanifold Framework for Log-Euclidean Metric Learning on Symmetric Positive Definite Manifolds

Orthogonal Single-Target Tracking

Deep Illumination-aware Dehazing with Low-light and Detail Enhancement

SphereGAN: Sphere Generative Adversarial Network Based on Geometric Moment Matching and Its Applications

Visual Tracking by Adaptive Continual Meta-Learning

Deep Plane Sweep Generative Adversarial Network for Consistent Multi-View Depth Estimation

Method for tracking object on visual

Nose pattern recognition system and method based on deep learning

System and Method for Data Processing using Sphere Generative Adversarial Network Based on Geometric Moment Matching

System and Method for Log Euclidean Metric Learning using Riemannian Submanifold Framework on Symmetric Positive Definite Manifolds

Method and apparatus for object recognition based on visible light and infrared fusion image

System and Method for Improving Low Light Level Image Using Generative Adversarial Network

Method and Apparatus For Deep-Learning Using Smoothing

Evaluation Method for Unsupervised 3D Point Cloud Generation

Automatic Abnormal Diagnosis Device and Method in Plant Process Design Using Deep Neural Network

DALE: Dark Region-Aware Low-light Image Enhancement

System and Method for Generating Highly Dense 3D Point Clouds using Wasserstein Distribution

Self-Distillation Dehazing

Pixel-wise Wasserstein Autoencoder for Highly Generative Dehazing

Improvement of Visual Tracking Using Adversarial Attack

Method And Computer Program For Performing Visual Tracking Using Wasserrstein Distance

Probabilistic model-based label self-enhancement system and method

Machine learning data preprocessing system and method for improving object detection performance

Uncertainty calibration system and method applying energy-based instance-specific scaling

Deep Metric Learning Method for Similarities between Images Introducing Class-agnostic Features

Apparatus And Method For Generating 3D Point Map From RGB Image Information

Apparatus And Method For Time-Series Anomaly Detection

Efficient Music Generation Apparatus and Method From a Single Image Using Diffusion-Based Representations

Method and Apparatus for Classifier Initialization using Neural Collapse-informed Direction and Perturbation Injection

Seongsoo Jeon

Junseok Kwon

Professor

Computer Vision Machine Learning Lab (CVML@CAU)

교수님 소개

Seongsoo Jeon

연구실 소개

연구실 키워드

알럼나이가 간 회사

연구실 지원 방법

대우 조건

연도별 논문 인용수

저널 & 학회 논문 출판 수

연구실 논문 목록

Calibration-Free Multi-Object Tracking via Scale-Invariant Topological Depth and Adaptive Uncertainty Gating

WildTalker ∞ : Pushing the Limits of 3D Talking Portrait Synthesis in Unconstrained Environments

Spectral-guided Physical Dynamics Distillation

연구실 보유 특허

대학별 입시끈

우리 연구실 모집

교수님 소개

Seongsoo Jeon

연구실 키워드

알럼나이가 간 회사

연구실 지원 방법

대우 조건

연도별 논문 인용수

저널 & 학회 논문 출판 수

연구실 논문 목록

Calibration-Free Multi-Object Tracking via Scale-Invariant Topological Depth and Adaptive Uncertainty Gating

WildTalker ∞ : Pushing the Limits of 3D Talking Portrait Synthesis in Unconstrained Environments

Spectral-guided Physical Dynamics Distillation

연구실 보유 특허

최근 3년간 연구실적 높은 연구실

대학별 입시끈

우리 연구실 모집