Scotoma Carving Data Sets August2021
The 3 datasets below each contain about 40 PowerPoint files, each showing the processing of one image. Each PowerPoint file contains an original image with a few fixation points marked, cutout sections with overlaid scotomas, and the carved images. There are 3 types of carving/scotoma combinations: :
- Partial Arcuate Scotoma (glaucoma), with for the effect of Eccentricity (Pelimodel) applied before carving
- Partial Arcuate Scotoma WITHOUT Pelimodel Applied
- Central Scotoma (disciform AMD) with Pelimodel applied before carving
The contents of each PowerPoint file (the processing of a single image), are:
- Two pages of contents description and parameters. Among those parameters are image name, image size, numbers of fixations, and type of scotoma used.
- Fixation points were selected manually to represent some of the most salient items in the image and positioning (enabling the entire scotoma to be in the frame). These are shown as yellow crosses on the original image (slide 3)
- The image was broken into sections surrounding each fixation point. If a selected fixation point was unsuitable due to positioning, it was eliminated from the set of selected points. Each section was sized in pixels to be a power of 2 for ease of processing with the Pelimodel and given a width of 60 degrees visual angle.
- On each section, the scotoma outline was overlayed in yellow, and yellow boxes were drawn delineating the cut section and the expected size of the section after processing.
- The final slides show the processed sections with the fixation position centered in the image and marked.
In inspecting the output images, it is difficult to note the image details that were carved out without prior examination and comparison with the outlined scotoma image. Removal of the regions within the scotomas results in barely distinguishable distortions around the carved out areas (for example, straight lines becoming curved – see the Stonehenge image: pexels-john-nail-1448136_VHFxOffset_(300,0))_arcuate.pptm). Another type of distortion is visible in faces (see image pexels-fauxels-3182812_VHFxOffset_(300,0))_arcuate), but even there, the distortions are hard to discern without comparing the processed to the original image.