|
|||
|
File
Formats and Directory Structure Image Sets: |
Ladder
Number Of Images: Total of 1159 images taken in 7
rooms: classroom (170), conference room (183), cubicle (174), apartment (174),
kitchen (163), lab (123), warehouse (171) Image Size: 1600 x 1200 Comments: This is a large set of
real images of a ladder at arbitrary out-of-image-plane rotation in 7
different prototypical environments : classroom, conference room, office, lab,
apartment, warehouse, kitchen (the example image shows a typical ÒclassroomÓ
image). For each image, the camera was approximately 3m away from the objects
in the scene; the elevation of the camera varied between 1.6m and 1.75m; the
camera position across the set of all images of a particular scene covered
about 60 degrees of rotation with respect to the scene objects in the plane
parallel to the floor. The camera was moved between each view, and once every
five views the ladder was rotated to an arbitrary angle with respect to the
ground and the poses and configurations of clutter objects were randomly
modified. The depth of the ladder with respect to the camera varied by a
total of approximately 20% across all views. This data set provides a testbed for evaluating how
shape-based recognition algorithms fare across a variety of realistic
environments. Vista Parameters: (sigma, lowThreshold,
highThreshold,minimumEdgeLength, accuracy, granularity, magnitude) = (1.5,10,50,25,2.0,4.0,20.0)
(lab, kitchen, conf. room) (2,10,50,25,2.0,4.0,20.0)
(cubicle, apartment) (3,10,50,25,2.0,4.0,20.0)
(classroom, warehouse) Download: (copyright) All rooms: [ladder.tar.gz , 522
MB ]. Individual
rooms: [ ladder_classroom.tar.gz
, 75 MB ][ ladder_conf.tar.gz
, 80 MB ] [ ladder_cubicle.tar.gz
, 81 MB ][ ladder_home.tar.gz
, 79 MB ] [ ladder_kitchen.tar.gz
, 74 MB ][ ladder_lab.tar.gz
, 53 MB ] [ ladder_warehouse.tar.gz
, 79 MB] |
||
WORD: Wiry
Object Recognition Database
