Project Topics

CSC I6716 Section 2RR -- COMPUTER VISION, Fall 2007

Note: a writing report is needed; please hand it to me in class right before your final presentation.

 

1.    Image decoding from Bayer pattern sources.

§     Double-check your basic code to make sure everything is all right in Assignment 1. This will be your baseline.

§     Improvements: try to use structures, e.g. edges; or better interpolation methods

§     Compare the improved results with the baseline. You will also need to do an analysis, for example, using histogram technique on the difference image.

2.    Sketch generation using edge detection methods in Assignment 2.

§     Gradient image using Sobel, etc.

§     Adaptive thresholding to cope with different spatial details and/or dynamic ranges at different portions of an image.

§     Contour extraction by Hough Transform (see lecture notes!) or edge tracking. (check the textbook or online materials).

3.    Stereo matching using the fundamental matrix you have obtained in Assignments 4.

§     Select matching primitives - e.g. square window, corner, edge, etc.

§     Determine matching criterion (SAD, SSD, etc.).

§     Perform matching (try to develop an efficient algorithm); please do a time analysis

§     Visualization to show 3D or disparities

4.    Self-selected Topics if you like, but you will need to let me know in a week.