An Image Processing Library Based on Abstract Image Data-Types in C++

The paper presents a library for image processing based on abstract image data-types. The library finds a balance between programmability and efficiency of implementation of image processing operations. Operations are designed using abstract concepts for images and pixel manipulation to be independent of the image data-type. Operations may be optimized in an iterative way by replacing abstract concepts with data-type specific operations, reaching the level of efficiency of operations designed from scratch for a specific image data-type.