Fruit and Texture Images

This page contains source images and data files used in the paper, A Comparison of Measures for Detecting Natural Shapes in Cluttered Backgrounds, by Lance Williams and Karvel Thornber. This paper will be presented at the Fifth European Conf. on Computer Vision (ECCV '98), in Freiburg, Germany.

Texture Images *

  1. Bark.0012.tif
  2. Brick.0000.tif
  3. Fabric.0000.tif
  4. Leaves.0000.tif
  5. Sand.0005.tif
  6. Stone.0003.tif
  7. Terrain.0000.tif
  8. Water.0004.tif
  9. Wood.0000.tif

Cropped Texture Images

  1. bark.tif
  2. brick.tif
  3. fabric.tif
  4. leaves.tif
  5. sand.tif
  6. stone.tif
  7. terrain.tif
  8. water.tif
  9. wood.tif

Cropped Texture Edges **

  1. bark.edge
  2. brick.edge
  3. fabric.edge
  4. leaves.edge
  5. sand.edge
  6. stone.edge
  7. terrain.edge
  8. water.edge
  9. wood.edge

Color Images of Fruits and Vegetables ***

  1. avocados1.tif.Z
  2. banana1.tif.Z
  3. lemon1.tif.Z
  4. peach1.tif.Z
  5. pear1.tif.Z
  6. redonion1.tif.Z
  7. sweet_potato1.tif.Z
  8. tamarillo1.tif.Z
  9. yellowapple1.tif.Z

Fruit and Vegetable Edges

  1. avocados1.edge
  2. banana1.edge
  3. lemon1.edge
  4. peach1.edge
  5. pear1.edge
  6. redonion1.edge
  7. sweet_potato1.edge
  8. tamarillo1.edge
  9. yellowapple1.edge
* From MIT Media Lab.
** The first line of the .edge file contains a single integer which is equal to the number of edges. Edges are stored one per line. The three fields on each line contain the y-location, x-location, and direction of each edge (in degrees).
*** From NEC Research Institute.