***********************************************************************
Test images for line/edge detection, snake detection, and texture seg-
mentation.
All are raw image files with 256x256 pixels, one byte per pixel.
 ls -l will show 65536 bytes (256^2)
The file "head" is an ASCII header file with only 4 lines.
Example: convert raw to gif format in order to display the file:
 cat head snake.imf > snake.gif

snake.imf is for testing continuity models = visual perception
  (snakes formed by Gabor functions)
ledge.imf is for testing line/edge detection with complex crossings
  (circular centre and ring with two faint diagonal lines)
mit.imf is a standard test image for edge detection and segmentation
  (it's an image of a bank I think, not an MIT building)
patch*.imf are for testing texture segmentation methods

Note: these images have been used for developing vision models which
are based on V1 simple and complex cells, i.e. complex Gabor filters
tuned to different orientations and scales.

***********************************************************************
patch*.imf are 5 images for testing texture classification or segmentation.
Each image contains 7 textures in 16 patches, such that all texture
combinations share a common boundary at least once. Horizontal boundaries
are straight while vertical ones follow a random walk. Texture amplitudes 
are normalised between 0 and 255, either by max and min values or by taking
the mean and variance (clipping tails).

Image:       Type:
patch1.imf - aerial textures
patch2.imf - aerial
patch3.imf - brodatz (from the Phil Brodatz 1966 "Textures" book)
patch4.imf - aerial
patch5.imf - synthetic

Most images have been used in publications on feature extraction and image
segmentation, so that reference results are available:

patch5:
du Buf, J.M.H., Spann, M.\ and Kardan, M.\ (1990) Texture feature
  performance for image segmentation. Pattern Recognition
  \underline{23}, pp.\ 291-309.

patch1,patch2,patch3:
du Buf, J.M.H.\ (1992) Abstract processes in texture discrimination.
  Spatial Vision \underline{6}, pp.\ 221-242.

patch1,patch2,patch3:
Big\"un, J.\ and du Buf, J.M.H.\ (1993) N-folded symmetries by
  complex moments in Gabor space and their application to unsupervised
  image segmentation. IEEE Trans.\ Pattern Analysis and Machine Intelligence,
  PAMI \underline {16}, pp.\ 80-87.

Good luck! Hans du Buf
************************************************************************

