| CARVIEW |
I'm a Senior Research Scientist/Engineer at Adobe in Seattle, specializing in controllable image and video generation with a focus on 3D-centric controls. My research bridges computer vision, machine learning, and computational photography to build systems that give creators precise control over visual content.
My work spans 3D vision (stereo, depth estimation, camera calibration), computational photography (burst processing, reflection removal, defocus synthesis), and video processing (view synthesis, depth estimation, scene-aware effects). I'm particularly interested in how 3D understanding can improve generative models and creative tools.
I earned my PhD from Cornell University (2016) under Noah Snavely and Kavita Bala, where I developed methods for analyzing large-scale photo collections. Before joining Adobe, I spent five years at Facebook (2016-2021) on Computational Photography and XR Tech.
Note: Publications prior to 2021 are under the name Kevin Matzen.
Publications
SIGMA-GEN: Structure and Identity Guided Multi-subject Assembly for Image Generation
Removing Reflections from RAW Photos
Fast View Synthesis of Casual Videos with Soup-of-Planes
VideoDoodles: Hand-Drawn Animations on Videos with Scene-Aware Canvases
Perspective Fields for Single Image Camera Calibration
A Practical Stereo Depth System for Smart Glasses
One Shot 3D Photography
Consistent Video Depth Estimation
VPLNet: Deep Single View Normal Estimation With Vanishing Points and Lines
GeoStyle: Discovering Fashion Trends and Events
Synthetic Defocus and Look-Ahead Autofocus for Casual Videography
DeepMVS: Learning Multi-view Stereopsis
Low-cost 360 stereo photography and video capture
StreetStyle: Exploring world-wide clothing styles from millions of photos
All Weather Perception: Joint Data Association, Tracking, and Classification for Autonomous Ground Vehicles
NYC3DCars: A Dataset of 3D Vehicles in Geographic Context
No publications found matching the selected filter.