Donggon Jang

I am a Ph.D. candidate at the BREIL, part of the School of Electrical Engineering at KAIST. My PhD advisor is Dae-Shik Kim.

I am broadly interested in the intersection of computer vision and deep learning. My current research focuses on topics such as multimodal large language models (MLLMs), semantic segmentation, domain adaptation, adversarial attack, and image restoration. I am particularly curious about how these techniques can be combined to build more robust and generalizable vision systems.

  GitHub  /  Google Scholar  /  LinkedIn

Research

I'm interested in computer vision, multimodal large language model, domain adapation, adversarial attack, and image restoration. My current work mainly focuses on multimodal reasoning segmentation and safety of multimodal large language models.

MMR: A Large-scale Benchmark Dataset for Multi-target and Multi-granularity Reasoning Segmentation

Energy-Based Domain Adaptation Without Intermediate Domain Dataset for Foggy Scene Segmentation

Maximizing discrimination capability of knowledge distillation with energy function

Temporally averaged regression for semi-supervised low-light image enhancement

Accurate hippocampus segmentation based on self-supervised learning with fewer labeled data

Strengthening the transferability of adversarial examples using advanced looking ahead and self-cutmix

Learning Color Representations for Low-Light Image Enhancement

Unsupervised Image Denoising with Frequency Domain Knowledge

Design and source code from Jon Barron's website