Picture for Heesung Kwon

Heesung Kwon

Exploring the Impact of Synthetic Data for Aerial-view Human Detection

May 27, 2024
Viaarxiv icon

Diversifying Human Pose in Synthetic Data for Aerial-view Human Detection

May 24, 2024
Viaarxiv icon

TK-Planes: Tiered K-Planes with High Dimensional Feature Vectors for Dynamic UAV-based Scenes

May 04, 2024
Viaarxiv icon

UAV-Sim: NeRF-based Synthetic Data Generation for UAV-based Perception

Oct 25, 2023
Viaarxiv icon

Novel Categories Discovery from probability matrix perspective

Add code
Jul 07, 2023
Figure 1 for Novel Categories Discovery from probability matrix perspective
Figure 2 for Novel Categories Discovery from probability matrix perspective
Figure 3 for Novel Categories Discovery from probability matrix perspective
Figure 4 for Novel Categories Discovery from probability matrix perspective
Viaarxiv icon

NEV-NCD: Negative Learning, Entropy, and Variance regularization based novel action categories discovery

Add code
Apr 14, 2023
Figure 1 for NEV-NCD: Negative Learning, Entropy, and Variance regularization based novel action categories discovery
Figure 2 for NEV-NCD: Negative Learning, Entropy, and Variance regularization based novel action categories discovery
Figure 3 for NEV-NCD: Negative Learning, Entropy, and Variance regularization based novel action categories discovery
Figure 4 for NEV-NCD: Negative Learning, Entropy, and Variance regularization based novel action categories discovery
Viaarxiv icon

Progressive Transformation Learning For Leveraging Virtual Images in Training

Add code
Nov 03, 2022
Figure 1 for Progressive Transformation Learning For Leveraging Virtual Images in Training
Figure 2 for Progressive Transformation Learning For Leveraging Virtual Images in Training
Figure 3 for Progressive Transformation Learning For Leveraging Virtual Images in Training
Figure 4 for Progressive Transformation Learning For Leveraging Virtual Images in Training
Viaarxiv icon

Negative Samples are at Large: Leveraging Hard-distance Elastic Loss for Re-identification

Jul 20, 2022
Figure 1 for Negative Samples are at Large: Leveraging Hard-distance Elastic Loss for Re-identification
Figure 2 for Negative Samples are at Large: Leveraging Hard-distance Elastic Loss for Re-identification
Figure 3 for Negative Samples are at Large: Leveraging Hard-distance Elastic Loss for Re-identification
Figure 4 for Negative Samples are at Large: Leveraging Hard-distance Elastic Loss for Re-identification
Viaarxiv icon

A Multi-purpose Real Haze Benchmark with Quantifiable Haze Levels and Ground Truth

Jun 13, 2022
Figure 1 for A Multi-purpose Real Haze Benchmark with Quantifiable Haze Levels and Ground Truth
Figure 2 for A Multi-purpose Real Haze Benchmark with Quantifiable Haze Levels and Ground Truth
Figure 3 for A Multi-purpose Real Haze Benchmark with Quantifiable Haze Levels and Ground Truth
Figure 4 for A Multi-purpose Real Haze Benchmark with Quantifiable Haze Levels and Ground Truth
Viaarxiv icon

Exploring Cross-Domain Pretrained Model for Hyperspectral Image Classification

Apr 07, 2022
Figure 1 for Exploring Cross-Domain Pretrained Model for Hyperspectral Image Classification
Figure 2 for Exploring Cross-Domain Pretrained Model for Hyperspectral Image Classification
Figure 3 for Exploring Cross-Domain Pretrained Model for Hyperspectral Image Classification
Figure 4 for Exploring Cross-Domain Pretrained Model for Hyperspectral Image Classification
Viaarxiv icon