Index of /~chandra/689/Spring06/DATA

[ICO]NameLast modifiedSizeDescription

[DIR]Parent Directory   -  
[IMG]face.left.pgm 12-Jan-2005 17:25 96K 
[IMG]face.right.pgm 12-Jan-2005 17:25 96K 
[VID]ref.G9_1621_x512.y512 11-Apr-2006 19:54 512K 
[VID]ref.G9_1634_x512.y512 11-Apr-2006 19:54 512K 
[   ]ref.G9_1650_x512.y512 11-Apr-2006 19:54 512K 
[   ]test.G8_1621_x512.y512 11-Apr-2006 19:54 512K 
[VID]test.G8_1634_x512.y512 11-Apr-2006 19:54 512K 
[   ]test.G8_1650_x512.y512 11-Apr-2006 19:54 512K 
[IMG]tsukuba.left.pgm 12-Jan-2005 17:25 108K 
[IMG]tsukuba.right.pgm 12-Jan-2005 17:25 108K 
[IMG]tsukuba.truedisp.rig..>12-Jan-2005 17:25 108K 

STEREO PAIRS: In each case, the images are given in 2 files, named *.left.XXX and *.right.XXX. (XXX indicates the image format) ------------- tsukuba.left.pgm tsukuba.right.pgm: These are the famous tsukuba images for stereo (more can be obtained from http://cat.middlebury.edu/stereo/. I strongly recommend to visit this site and test your algorithms). The central image is chosen as the right image for which the ground truth is given. The disparity range is about [5,30] face.left.pgm face.right.pgm: These are stereo images from Jim Little (UBC). The disparity in the left image is about [50, 80] relative to the right image. The images appear to be sideways because the cameras were oriented vertically relative to the face, but the rows match up accurately. ------------- NEW!! Hurricane Luis Data There are 3 stereo pairs (left image: ref.G9_XXXX_x512.y512, right image: test.G8_XXXX_x512.y512). XXXX indicates time in UTC, taken by Goes-8 or Goes-9 Geo-stationary satellite. (you will be performing stereo between images having same time) The data type is unsigned short, size is 512X512. You should get pretty good depth(disparity) results with this data.