CSE/EE 486:  Fundamentals of Computer Vision


      Computer Project Report #0 :

      Introduction to the Lab (Cantata/xv)

      Anirudh Modi

      Date: 01/20/1999

      1. Objectives:
        To get acquainted with image capturing and use of Cantata and xv.

      2. Methods:
        The following was done in the course of this project:
        1. The program SunVideo was executed on the VORLON workstation connected to a video camera.
        2. A gray-scale image of myself was captured using the software and stored as a Sun Raster file named modi_orig.rs.
        3. The captured image (which was of size 320x240) was loaded under xv. An approximate square window was selected using the mouse and the image was cropped and then resized to 256x256 and saved back as a Sun Raster file named modi.rs.
        4. Cantata was now started and a glyph for reading the captured image was placed in the workspace. This was done by selecting INPUT SOURCES->Input Data File->User Defined and then picking up modi.rs as the User Specified File.
        5. Another glyph was added to convert the Sun Raster image to VIFF image by selecting CONVERSIONS->Standard File Format->Raster to VIFF.
        6. The image was then altered by placing the following glyphs (which acted as filters) in the order mentioned below:
          1. IMAGE PROCESSING->Spatial Filters->Sobel with the default parameters.
          2. IMAGE PROCESSING->Spatial Filters->Speckle Removal with the ``Number of Iterations'' increased to 5.
        7. The last glyph was finally connected to two different glyphs simultaneously; one for displaying the image on the screen (OUTPUT->Display Image->Display With Update), and the other for storing the resultant VIFF image back into a Sun Raster image modi1.rs (CONVERSIONS->Standard File Format->VIFF to Raster).

      3. Results:
        The resultant images can be found in the Appendix below. The input image (Figure 2) was obtained after cropping and scaling the captured image (Figure 1) using xv. The final image (Figure 3) was obtained after applying the Sobel and Speckle Removal filters to the input image.

      4. Conclusions:
        Capturing images using the digital video camera in the lab was found to be really easy. After a brief exposure, it was concluded that Cantata is indeed a very powerful and versatile software package for image processing. The ability to interactively create complicated procedures to manipulate images in a very short time was impressive. xv was found to be a quick and handy tool for converting between various commonly used image formats.

      5. Appendix:
        1. Images:

          Captured image
          Figure 1: Captured image (          modi_orig.rs)
          Original After applying filters
          Figure 2: Input image (modi.rs) Figure 3: Output image (modi1.rs)


        2. The file project0.wksp contains the Cantata workspace settings to execute the above procedure.