A simple computervision library in c++

Workflow:

• Load image as PPMImage
• Apply edge filter

To Do:

• Add tests for the normalise method.
• Split the pixel class into two types: rgb and mono
  • rgb can then also have methods for getting things like YUV values to allow converting to mono
  • The PixelGrid class can then be used for all processing, convolutions etc.
• Pseudo colorise processed images by resizing the original to the new size and then using the color map scaled by the mono intensity.
• Write unit tests for modules
  • Get some simple images and check that the convolutions work
• Do a full implementation of canny edge detection

Goals:

• Loading in simple image files
• Displaying images.
• Applying Convolution filters to images.
• Write a script which will read in a file containing instructions:
• specify and image, specify a set of kernels, specify an output file name
• load the image and convolve with the filters and save the output.
• Generate key filter kernels at given sizes.
• ??
• Blurring and downscaling images
• Interpolating and upscaling images
• ??
• Edge detection
• ??
• Feature detection (Something like SIFT)
• ??