T
Thomas Brox
Researcher at University of Freiburg
Publications - 353
Citations - 127470
Thomas Brox is an academic researcher from University of Freiburg. The author has contributed to research in topics: Segmentation & Optical flow. The author has an hindex of 99, co-authored 329 publications receiving 94431 citations. Previous affiliations of Thomas Brox include Dresden University of Technology & University of California, Berkeley.
Papers
More filters
A system for marker-less motion capture
TL;DR: This paper presents a silhouette based human motion capture system, which contains silhouette extraction based on level sets, a correspondence module, which relates image data to model data and a pose estimation module, and performs a comparison of the motion estimation system with a marker based tracking system.
Posted Content
Object Detection, Tracking, and Motion Segmentation for Object-level Video Segmentation
Benjamin Drayer,Thomas Brox +1 more
TL;DR: An approach for object segmentation in videos that combines frame-level object detection with concepts from object tracking and motion segmentation and provides an accurate, temporally consistent segmentation of each object is presented.
Posted Content
Parting with Illusions about Deep Active Learning
TL;DR: This work re-implement various latest active learning approaches for image classification and evaluate them under more realistic settings and realistically assess the current state of the field and propose a more suitable evaluation protocol.
Book ChapterDOI
High accuracy optical flow serves 3-d pose tracking: exploiting contour and flow based constraints
TL;DR: In this article, the authors proposed to use two complementary types of features for pose tracking, such that one type makes up for the shortcomings of the other, and they employed the optic flow to compute additional point correspondences.
Posted Content
3D Human Pose Estimation in RGBD Images for Robotic Task Learning
TL;DR: This work proposes an approach to estimate 3D human pose in real world units from a single RGBD image and shows that it exceeds performance of monocular 3D pose estimation approaches from color as well as pose estimation exclusively from depth.