Scholar
Kangning Liu
Google Scholar ID: F3F2qAkAAAAJ
Adobe Inc
Citations & Impact
Citations
721
H-index
10
i10-index
10
Publications
20
Co-authors
0
Contact
Publications
Resume (English only)
Academic Achievements
- Multiple Ph.D.-era research projects successfully transferred into Adobe products.
- Contributed to research under imperfect supervision, including:
- - Uncertainty-aware fine-tuning of segmentation foundation models (SUM)
- - Noise-resilient deep segmentation (ADELE)
- - Weakly supervised segmentation (GLAM)
- - Unsupervised/self-supervised learning (ItS2CLR)
- - Video analysis (StrokeRehab)
- - Video synthesis (UVIT & Controllable Face Video Synthesis)
Research Experience
- Research Scientist at Adobe, San Jose, California, USA (April 2024 – present):
- - Developed cutting-edge segmentation technologies deployed in Adobe Photoshop’s selection tools, Firefly’s background removal, and segmentation enhancements in Lightroom and Adobe Express.
- - Enhanced grounding capabilities of multimodal large language models with advanced reasoning and fine-grained understanding for generative AI data creation, reward modeling, and image editing.
- Research Intern at Adobe, San Jose, California, USA (May 2023 – Nov 2023), advised by Dr. Brian Price, Dr. Jason Kuen, Dr. Yifei Fan, Dr. Zijun Wei, Luis Figueroa, and Markus Woodson:
- - Engaged in a large-scale segmentation initiative using over 100 million diverse images, with a model-in-the-loop data engine and close collaboration across Adobe teams.
- - Reproduced SAM’s large-scale multi-node training and interactive segmentation, further improving accuracy and efficiency with potential product impact.
- - Initiated a human-centric feedback loop design.
Co-authors