ELEC4622 - Multimedia Image Processing

Project1

1. Reduce the image by the factor of 2/5 with windowed sinc filters
   • Direct filtering with 2D Kernels
   • separability as an important basis for minimizing the complexity
2. Expand the image by the factor of 5/2 with bi-linear interpolation
3. Reduce the image by the factor of 5/2 with windowed sinc filters
   • Direct filtering with 2D Kernels
   • separability as an important basis for minimizing the complexity
4. Compare two images in terms of MSE and PSNR and write out the difference image

Project2

1. Generate a bi-level image given with the threshold
2. Morphological erosion
3. Morphological erosion with roughly circular structuring sets of various radius values r
4. Morphological opening
5. Morphological closing

Project3

1. Global translational motion estimator that uses a fixed set of keypoints
2. Motion compensated using bilinear interpolation
3. Keypoint detection with the Harris corner detector