DongGoo Kang

DongGoo Kang

Vision Part Lead @ Smilegate

Computer Vision · Graphics AI · Vision-Language Models

I am a Ph.D. with extensive experience in Computer Vision, Graphics AI, and Vision-Language Models. My work focuses on bringing AI into the 3D world—bridging perception and generation to create intelligent, immersive digital environments.

I currently lead the Vision Part at Smilegate, building large-scale 2D/3D generative vision systems for games. I received my Ph.D. in AI Imaging from Chung-Ang University, advised by Prof. Joonki Paik. Feel free to reach out by email.

News

Articles

Experience

2026 – Present

Vision Part Lead · Smilegate

Large-scale 2D/3D vision serving for in-game AI.

  • Designed and scaled production serving systems for 2D/3D generation.
  • Managed H200 GPU clusters with TensorRT optimization, quantization, and low-latency inference pipelines.
  • Handled 7 TB+ datasets and model deployment in closed-network environments.
  • Collaborated with Unreal Engine 5 teams on in-game 2D/3D asset integration.
2025 – 2026

Research Scientist · Smilegate

AI Center — Creative AI Team. 3D asset generation.

2024

Research Intern · NC AI (NCSOFT)

Graphics AI Lab — AI Texturing Team. Advancing 3D texturing and UV mapping.

  • Applied latent video generation models to overcome the limitations of traditional texturing.
  • Integrated ControlNet into a vid2vid framework for geometry-aware texturing.
  • Used positional encoding to improve accuracy and consistency in UV inpainting.

Education

2021 – 2025

Ph.D. in AI Imaging · Chung-Ang University

GPA 4.5 / 4.5. Advisor: Prof. Joonki Paik.
Image Processing and Intelligent System (IPIS) Lab.
Thesis: A Study on Zero-shot 3D Texture Synthesis Using Geometry-aware Diffusion and Temporal Video Models.

2018 – 2020

M.S. in AI Imaging · Chung-Ang University

GPA 4.39 / 4.5. Advisor: Prof. Joonki Paik.
Image Processing and Intelligent System (IPIS) Lab.
Thesis: Salient Object Detection Using Bilateral Attention Network With Dice Coefficient Loss.

Honors & Awards

Publications

175 citations, h-index 7 · full list on Google Scholar. Author name in bold.

  1. Enhancing video frame interpolation with region of motion loss and self-attention mechanisms.
    Y Kim, S Kwon, D Kang, H Lee, J Paik. Neurocomputing, 2025.
  2. Structure-preserving zero-shot image editing via stage-wise latent injection in diffusion models.
    D Jeong, D Kang, J Park, H Lee, J Paik. arXiv preprint, 2025.
  3. Analyzing South Korea's defense challenges with Defense Innovation 4.0 through LDA and BERT techniques.
    D Park, D Kang, J Paik. IEIE Transactions on Smart Processing & Computing, 2025.
  4. Cosine similarity-guided knowledge distillation for robust object detectors.
    S Park, D Kang, J Paik. Scientific Reports, 2024.
  5. Real-time human group detection and clustering in crowded environments using enhanced multi-object tracking.
    H Lee, D Kang, H Park, S Park, D Jeong, J Paik. IEEE Access, 2024.
  6. VLM-HOI: Vision language models for interpretable human-object interaction analysis.
    D Kang, D Jeong, H Lee, S Park, H Park, S Kwon, Y Kim, J Paik. European Conference on Computer Vision (ECCV), 2024.
  7. Enhanced human–object interaction detection via maximum IoU partitioning and chunk block attention.
    H Lee, D Kang, H Park, Y Kim, S Kwon, J Paik. IEEE Access, 2024.
  8. Generating diverse image variations with diffusion models by combining intra-image self-attention and decoupled cross-attention.
    D Jeong, D Kang, J Park, J Paik. IEEE Access, 2024.
  9. Multiple object detection method and apparatus.
    JK Paik, SW Park, DG Kim, DG Kang. US Patent 11,816,881, 2023.
  10. A review on few-shot learning for medical image segmentation.
    Y Kim, D Kang, Y Mok, S Kwon, J Paik. Int. Conf. on Electronics, Information, and Communication (ICEIC), 2023.
  11. Bidirectional meta-Kronecker factored optimizer and Hausdorff distance loss for few-shot medical image segmentation.
    Y Kim, D Kang, Y Mok, S Kwon, J Paik. Scientific Reports, 2023.
  12. Action recognition using multi-stream 2D CNN with deep learning-based temporal modality.
    K Kang, S Park, H Park, D Kang, J Paik. IEEE Int. Conf. on Consumer Electronics (ICCE), 2023.
  13. Video quality assessment system using deep optical flow and Fourier property.
    D Kang, Y Kim, S Kwon, H Kim, J Kim, J Paik. IEEE Access, 2023.
  14. Multiple-scale fire detection based on one-stage deep learning model.
    I Lee, Y Lo, D Kang, H Park, J Paik. Int. Conf. on Electronics, Information, and Communication (ICEIC), 2022.
  15. Coarse to fine: Progressive and multi-task learning for salient object detection.
    D Kang, S Park, J Paik. Int. Conf. on Pattern Recognition (ICPR), 2021.
  16. SdBAN: Salient object detection using bilateral attention network with dice coefficient loss.
    D Kang, S Park, J Paik. IEEE Access, 2020.
  17. Real-time robust object detection using an adjacent feature fusion-based single shot multibox detector.
    D Kim, S Park, D Kang, J Paik. IEIE Transactions on Smart Processing & Computing, 2020.
  18. Uniformity attentive learning-based Siamese network for person re-identification.
    D Jeong, H Park, J Shin, D Kang, J Paik. Sensors, 2020.
  19. Improved center and scale prediction-based pedestrian detection using convolutional block.
    D Kim, S Park, D Kang, J Paik. IEEE Int. Conf. on Consumer Electronics (ICCE-Berlin), 2019.
DongGoo Kang