Jaedong Hwang
Research
CVPR 2021

Exemplar-Based Open-Set Panoptic Segmentation Network

  1. †Seoul National University
  2. ‡Adobe Research

*This work was done during an internship at Adobe Research

Project page arXiv Code

Additional Qualitative Results

Images

Ground-Truths

Predictions

Image 1 Ground truth 1 Prediction 1 Image 2 Ground truth 2 Prediction 2 Image 3 Ground truth 3 Prediction 3 Image 4 Ground truth 4 Prediction 4 Image 5 Ground truth 5 Prediction 5 Image 6 Ground truth 6 Prediction 6 Image 7 Ground truth 7 Prediction 7 Image 8 Ground truth 8 Prediction 8

More research

Mid-Training Language Models on Raw VideoarXiv 2026
Sensor Geometry as a Flow-Matching Prior for Multi-Channel Brain SignalsarXiv 2026
B[FM]²: Brain Foundation Model via Flow Matching with SplitUNetNeurIPS 2026
Large Pre-Training Datasets Don’t Guarantee Robustness after Fine-TuningTMLR 2026
Learn Globally, Speak Locally: Multilingual Reasoning (M2A)arXiv 2025
Learn Globally, Speak Locally: Multilingual Reasoning (BRIDGE)arXiv 2025
Grid Cell-Inspired Fragmentation and Recall for Efficient Map BuildingTMLR 2024
Weakly Supervised Instance Segmentation by Deep Community LearningWACV 2021