Analysis and Recognition in Images and Video – Projects

OverviewTopicsNotes & MaterialProjectsResources
Click for Course Overview
Click for Course Topics
Click for Course Notes& Material
Click for Course Projects
Click for Course Resources

 

Project 

The goal of this project is to study a state-of-the-art topic. You can do this by either implementing a recent paper, or by coming up with your own idea of a cool application.

A good  way to collect ideas is by going over papers published in one of the leading computer vision conferences (ICCV,CVPR or ECCV). Here is a link to all the papers of recent years:
http://www.cvpapers.com/index.html

Project proposal

You must send me a project proposal  (at most 2 pages long) describing the project you’ve selected.  The project proposal should include the following:

  • Paper details: Title, author names, where was the paper published.
  • Abstarct: a short summary of the problem you wish to solve.
  • Background: why is ths problem relevant and interesting.
  • Approach: the solution direction you intend to adopt.
  • Experiments: what experiments you plan to do and on what data.
  • References: list of related papers.

Final submission

  1. Project summary

Your project summary should include the following:

    • Paper details: Title, author names, where was the paper published.
    • Abstarct: a short summary of the problem you solved.
    • Background: why is ths problem relevant and interesting.
    • Approach: the solution direction you adopted.
    • Experiments: what experiments you did and on what data.
    • References: list of related papers.
    • Conclusions: This should include the following:
      • What worked?
      • What didn’t work?
      • What was the simplest step?
      • What was most difficult?
      • Have you changed any of the initial decisions?
      • What would you have done differently if you had to do it again?
      • How do you expect this to improve your results?

 

  1. Code+data

Your submission should include:

    • A “help” description for each function.
    • Coments explaining non-trivial steps in the code.
    • runme.m function that will be called and the code will run
    • Data required to run the code.

 

Here are a few papers I have selected (no need to stick to this list)


Topic Paper
Fun with images Optimizing Content-Preserving Projections for Wide-Angle Images
Robert Carroll, Maneesh Agrawala, Aseem Agarwala
SIGGRAPH’09
Content-Preserving Warps for 3D Video Stabilization
Feng Liu, Michael Gleicher, Hailin Jin, Aseem Agarwala
SIGGRAPH’09
Filter Flow (PDF, supplemental material)Steven M. Seitz, Simon BakerICCV’09
Image matching SIFT flow: dense correspondence across different scenes
Ce Liu, Jenny Yuen, Antonio Torralba, Josef Sivic, William T. Freeman ECCV’08
Scale & Affine Invariant Interest Point Detectors
K Mikolajczyk & C. Schmid,  IJCV’04
Linear Solution to Scale and Rotation Invariant Object Matching
Hao Jiang, Stella X. Yu CVPR’09
Multi-camera action recognition Multi-Camera Activity Correlation Analysis
Chen Change Loy, Tao Xiang, Shaogang Gong
CVPR’09
Free viewpoint action recognition using motion history volumes
D Weinland, R Ronfard, E Boyer –
Computer Vision and Image Understanding, 2006
 Motion history volumes for free viewpoint action recognition
D Weinland, R Ronfard, E Boyer –
IEEE International Workshop on Modeling People and Human , 2005
Action recognition Recognizing Realistic Actions from Videos “in the Wild”
Jingen Liu, Jiebo Luo, and Mubarak Shah
CVPR’09
Actions in context (project page)
Marcin Marszalek, Ivan Laptev, Cordelia Schmid
CVPR’09
Video summarization Understanding Videos, Constructing Plots – Learning a Visually Grounded Storyline Model from Annotated Videos
Abhinav Gupta, Praveen Srinivasan, Jianbo Shi, Larry S. Davis
CVPR’09
Saliency detection
Image features Compact Signatures for High-Speed Interest Point Description and Matching (PDF)Michael Calonder, Vincent Lepetit, Pascal Fua, Kurt Konolige, James Bowman, Patrick MihelichICCV’09