CS 283: Computer Vision

Final Projects

Final projects are to be done individually. A list of suggested topics is below. These topics were chosen because they are directly related to course material, are interesting, and are approximately the right scope (most are linked to a particular research paper). You are free to propose a topic that is not on this list.

Deliverables. You are required to complete both a written report and an oral presentation. The oral component will be brief and (probably) in the form of a poster. The written report will be between 5 and 15 pages and include (at least) the sections: Introduction, Methods, Results, Conclusion, and References.

Timeline: 

	Tues, Nov. 20, 11:30am	Project proposals submitted through a drop-box on the Assignments page of the course website. This is a PDF file with two or three paragraphs that outline the motivation, the goal, the proposed method, and a plan for evaluation (including a source of data, if applicable).
	Tues, Dec. 11, 11:30am	Oral presentations. Lunch will be served.
	Wed, Dec. 12, 11:59pm	Final reports submitted in PDF form through a drop-box on the Assignments page of the course website.

Collaboration and third-party code. Final projects are to be done independently. That said, we encourage you to continue to make use of the class mailing list to help each other out and to identify and fix bugs. You are encouraged to make use of OpenCV. If you would like to use other third-party libraries, please consult the teaching staff first (we are likely to say yes.)

Some online Resources: 

• openCV: Open source computer vision library. See this wiki for documentation.
• CVonline: Community-contributed tutorials and information on many, many topics.
• Computer Vision Homepage: Links to software, databases, etc.
• Keith Price Bibliography: Annotated bibliography.
• Matlab Image Processing Toolbox: Useful documentation.

Example Topics

Radiometry

• Photometric stereo. (Can use the Extended Yale Face Database B, for example, or capture your own.) The surface normals and the light sources can both be recovered using a factorization algorithm not unlike structure from motion. This gives a reconstruction up to an unknown affine ambiguity, but this ambiguity can be reduced by enforcing integrability:

  A. Yuille, and D. Snow. "Shape and Albedo from Multiple Images using Integrability." CVPR. 2007.

  The remaining three-parameter ambiguity is known as the generalized bas relief ambiguity. One way of resolving it is to exploit the fact that faces are not Lambertian. See:

  A. Georghiades. "Incorporating the Torrance and Sparrow Model of Reflectance in Uncalibrated Photometric Stereo." ICCV. 2003.

• Inverse rendering using spherical harmonics. Lots of stuff in this paper, but one manageable piece is recovering the lighting environment given a single image of a known shape and known BRDF. Contact Prof. Zickler for data if interested.

D. Mahajan, R. Ramamoorthi, and B. Curless. "A Theory Of Frequency Domain Invariants: Spherical Harmonic Identities for BRDF/Lighting Transfer and Image Consistency." IEEE Trans. Pattern Analysis and Machine Intelligence. 2008.

Multi-view geometry

• Rotation-based camera calibration. Section 19.6 in Hartley and Zisserman and the original paper:

  R. I. Hartley. "Self-calibration from multiple views with a rotating camera." ECCV. 1994.

• Projective factorization and 3D reconstruction. Section 18.4 in Hartley and Zisserman and the original paper:

  P. Sturm, and W. Triggs. "A factorization based algorithm for multi-image projective structure and motion." ECCV. 1996.

  This will give cameras and points up to an unknown projective transformation of P^3. Methods for resolving this ambiguity are described in Chapter 19 of Hartley and Zisserman, and a particularly simple method is here:

  M. Pollefeys, R. Koch, and L. J. Van Gool. "A simple and efficient rectification method for general motion." ICCV. 1999.

Edges and lines

• Understanding and using image contours. Some efficient methods are:

  P. Felzenszwalb, and J. Schwartz. "Hierarchical Matching of Deformable Shapes." CVPR. 2007.

  P. Felzenszwalb, and D. McAllester. "A Min-Cover Approach for Finding Salient Curves." IEEE Workshop on Perceptual Organization in Computer Vision. 2006.

Color and Texture

• Texture synthesis and inpainting

  A. Efros, and W. Freeman. "Image quilting for texture synthesis and transfer." ACM Transactions on Graphics. 2001.

  A. Criminisi, P. Perez, and, K. Toyama. "Region filling and object removal by exemplar-based inpainting." IEEE Transactions on Image Processing. 2004.

• Structure and texture separation

  J. Aujol, G. Gilboa, T. Chan, and S. Osher. "Structure-texture image decomposition—modeling, algorithms, and parameter selection." International Journal of Computer Vision. 2006.

  Z. Farbman, R. Fattal, D. Lischinski, and R. Szeliski. "Edge-preserving decompositions for multi-scale tone and detail manipulation." ACM Transactions on Graphics. 2008.

  M. Kass, and J. Solomon. "Smoothed local histogram filters." ACM Transactions on Graphics. 2010.

• Highlight removal

  Assume the color of the light source is known, and compute an "specularity-free image" according to:

  Todd Zickler, Satya P. Mallick, David J. Kriegman, and Peter N. Belhumeur, "Color Subspaces as Photometric Invariants." Proc. IEEE Conference on Computer Vision and Pattern Recognition, June 2006.

  Then "put back" the diffuse color and texture using inpainting techniques like those described above.

Classifiers and recognition

• Set-based face recognition.

  T. Kim, J. Kittler, and R. Cipolla, "Discriminative learning and recognition of image set classes using canonical correlations." IEEE Transactions on Pattern Analysis and Machine Intelligence. 2007.

• Texture recognition using filterbanks.

  S. Grigorescu, N. Petkov, and P. Kruizinga. "Comparison of texture features based on Gabor Filters." IEEE Transactions on Image Processing. 2002.

Segmentation

• Interactive segmentation using graph-cuts.

  C. Rother, V. Kolmogorov, and A. Blake, "GrabCut - Interactive Foreground Extraction using Iterated Graph Cuts." ACM Transactions on Graphics. 2004.

• Segmentation using active regions.

  N. Paragios, and R. Deriche, "Geodesic Active Regions for Texture Segmentation." International Journal of Computer Vision. 2002.

Photograph editing

• Poisson image editing.

  P. Perez, M. Ganget, and A. Blake, "Poisson Image Editing." ACM Transactions on Graphics. 2003.

• Automatic collage creation.

  C. Rother, L. Bordeaux, Y. Hamadi, and A. Blake, "AutoCollage." ACM Transactions on Graphics. 2006.

S. Goferman, L. Zelnik-Manor, and A. Tal, "Context-Aware Saliency Detection." CVPR. 2007.