Some New Directions in Energy Minimization with Graph Cuts

Channel:
Microsoft Research
Subscribers:
344,000
Published on ● Video Link: https://www.youtube.com/watch?v=ui4IzfzD7D0


Duration: 1:20:01
855 views
4

Algorithms based on graph cuts have had a major impact on an important class of vision problems.  In these problems, which arise in applications such as stereo, every pixel must be assigned a label from some predefined set.  There is a known cost to assign a given label to any pixel, as well as a known cost for assigning different labels to adjacent pixels.  I will describe some recent work that addresses a broader class of problems, where the labels are not specified in advance, and where the cost of assigning a given label to any pixel must be determined.  This broader class of problems arises from applications such as multi-modal imaging, stereo imaging with unknown camera gain/bias, texture segmentation and layered motion segmentation. I will present some preliminary results that suggest that these problems can be solved by combining graph cuts with ideas from Expectation-Maximization and mutual information.



Other Videos By Microsoft Research


2016-09-05Not Even Wrong
2016-09-05Measurement and Monitoring in Wireless Sensor Networks
2016-09-05From TimeSync to EmStar: What's really hard in sensor networks?
2016-09-05JET: Join-Exit Tree for Time Efficient Contributory Group Key Management
2016-09-05Large-scale 3D Reconstruction from Video
2016-09-05Analysis of Multiagent Teams using Distributed POMDPs
2016-09-05Understanding and Designing for Physically Large Displays
2016-09-05Distinguishing Chambers of the Moment Polytope
2016-09-05Exceptions and side-effects in atomic blocks
2016-09-05Social Catalysts: enhancing communication in mediated spaces
2016-09-05Some New Directions in Energy Minimization with Graph Cuts
2016-09-05The Personal Rover Project: Robotics for the Masses
2016-09-05How can we apply the most powerful tools of our time to the search for an effective HIV Vaccine?
2016-09-05The Location Stack: Design, Implementation, and Use of Probabilistic Multi-sensor Location Systems
2016-09-05The Flickering Mind: The False Promise of Technology in the Classroom and How Learning Can Be Saved
2016-09-05Content Delivery in the Modern Internet
2016-09-05Random walk and random aggregation, derandomized
2016-09-05Enabling NASA's New Vision for Space Exploration through Human-Centered Intelligent Systems
2016-09-05Maximizing the Spread of Influence in a Social Network
2016-09-05Fixing the String Kernel - A Semi-Definite Programming Approach
2016-09-05Stable Marriage of Poisson and Lebesgue


Tags:
microsoft research