For a long time, the checkboard image has been the most popularly used for testing the filtering and antialiasing techniques in image warping/texture mapping. While people are still struggling with the checkboard image, it is by no means the most challenging one. You can be assured that the "Star" images -- the images with radial lines emitting from the images' centers -- will challenge almost all the filtering techniques by far exist. Even to generate these images is very challenging.
Here I have generated a group of Star images for people who have the same interest as me and are willing to take the challenge. Try to tilt the images just a bit and see what immediately disturbes you --- an artifact called aliasing. Even if your technique can handle a static image well, now try to continuously rotate the image, artifacts show up instantly, which is called temporary aliasing.
The images shown here are only a few that I have randomly generated. If you need any variant of the images, you are welcome to contact me at baoquan 'at' cs.umn.edu.
Here is a code that I wrote for demonstrating texture mapping using different filtering methods: bilinear interpolation (default, or press 'e'), mim-mapping (press '9'), and footprint-assembly (press '3). The code runs on unix (SUN). It accepts 'ppm' and 'sgi' formats. Here is the infamous 'checkboard' benchmark image. Other binary codes are available upon request.
Publications
Image 2
Image 3
Image 4