Experimental Results:
I tested and experimented my program with 3 datasets(including the one provided to us). The results are as shown below:-
  1. Data Set 1 : (Mosaic Example provided)
    The source images for this dataset were:
    View 0 View 1 View 2 View 3

    The results of mosaicing them are as shown below:
    (click on image to see full sized version)

    View 0 Mosaic View 1 Mosaic
    View 2 Mosaic View 3 Mosaic


    Changing Number of Control Points

    The following images are the same as above but now with 4 control points only:
    The images are seriously distorted when we use only 4 control points as can be seen in the following images.
    (click on image to see full sized version)

    View 0 Mosaic View 1 Mosaic
    View 2 Mosaic View 3 Mosaic


    Comparison of Image Blending Techniques - No Blending/Averaging/Feathering(Weighted Averaging)
    Blending the seams between images is very important to obtain good quality mosaics with smooth transitions at the edges. The following images illustrate this point by showing the same mosaics using :-
    1. No Image Blending
    2. Averaging
    3. Weighted Averaging (Feathering)

    (click on image to see full sized version)
    View 0 without blending View 0 with averaging View 0 with Feathering (weighted averaging)
    View 1 without blending View 1 with averaging View 1 with Feathering (weighted averaging)
    View 2 without blending View 2 with averaging View 2 with Feathering (weighted averaging)
    View 3 without blending View 3 with averaging View 3 with Feathering (weighted averaging)


    Illustrating Contrast Adjustment
    Contrast adjustment is necessary if the camera exposure is different for the images to be mosaiced. as this will result in significant intensity differences to show up in the mosaic. I have described the contrast adjustment approach in the implementation section of the report. To illustrate contrast adjustment , i intentionally changed the illumination of one of the images and then ran the mosaicing program with contrast adjustment turned on. The source image whose contrast was changed is as shown below:


    View 2 with increased brightness

    The images in the following table compare mosaicing with contrast adjustment versus no adjustment.
    (click on image to see full sized version)
    View 0 Mosaic without contrast adjustment View 0 Mosaic with contrast adjustment
    View 1 Mosaic without contrast adjustment View 1 Mosaic with contrast adjustment
    View 2 Mosaic without contrast adjustment View 2 Mosaic with contrast adjustment
    View 3 Mosaic without contrast adjustment View 3 Mosaic with contrast adjustment


  2. Data Set 2 (An Simple Planar Mosaic )

    To test the planar mosaicing , i took a single image and cut it into 2 parts using the paint program. Now i manually found correspondences using xv between the two parts and gave them as input to the mosaicing program. The images below show the two parts that were stitched together to create the planar mosaic from the split pieces.
    Click to see enlarged version
    First Image Second Image


    Stitching these two images gives the following result:-


  3. Data Set 3 (My Panaromic Mosaic )

    This is a 180 degree panaromic mosaic i constructed using images of my room. These are the individual images:-
    This is the panaroma constructed:-