Brian Price
Principal Scientist
Adobe Research
Email:
As a research scientist at Adobe, I research Computer Vision, develop new technologies, transfer them to Adobe products, and work with universities and PhD students through collaborations and internships. (See full bio here)
Research Interests: Computer Vision, Graphics, and Machine Learning, especially segmentation (interactive object selection, matting, panoptic/semantic segmentation, video segmentation), image synthesis, and document understanding.
Latest News
CVPR 2023
Jun 19, 2023
I'm at CVPR in Vancouver this week. I have three papers being presented. Check out my Papers page for details.
CVPR acceptances
Mar 1, 2023
I had three papers accepted to CVPR 2023. Watch my Papers page for these to appear as CVPR approaches.
ECCV 2022
Oct 25, 2022
It's ECCV this week. I have one paper, One-Triamp Video Matting. I also had an oral paper at the TiE: Text in Everything workshop. Check them out on my Papers page.
Recent Publications
ObjectStitch: Object Compositing With Diffusion Model
Yizhi Song, Zhifei Zhang, Zhe Lin, Scott Cohen, Brian Price, Jianming Zhang, Soo Ye Kim, Daniel Aliaga
CVPR 2023
[PDF]
GamutMLP: A Lightweight MLP for Color Loss Recovery
Hoang M. Le, Brian Price, Scott Cohen, Michael S. Brown.
CVPR 2023
[PDF]
Towards Open-World Segmentation of Parts
Tai-Yu Pan, Qing Liu, Wei-Lun Chao, Brian Price
CVPR 2023
[PDF]
Recent Patents
Generating Class-Agnostic Object Masks in Digital Images
Brian Price, Scott Cohen, Yinan Zhao.
US Patent 11587234. Feb. 21, 2023.
Utilizing interactive deep learning to select objects in digital visual media
Brian Price, Scott Cohen, Mai Long, Jun Hao Liew.
US Patent 11568627. Jan. 31, 2023.
Text Refinement Network
Zhifei Zhang, Xingqian Xu, Zhaowen Wang, Brian Price.
GB Patent 2600806. Dec. 28, 2022.
Technologies Highlights
Object-aware Refine Edge
A few years back I helped to advise on the creation of the Select and Mask workspace in Photoshop that allows users to select objects in images. Photoshop added a new mode that allows for improved extraction of hair for complex images. This mode is a follow on to our Deep Image Matting paper from CVPR 2017.