Seeing a Toy Crane (original) (raw)
Challenge for Vision: Seeing a Toy Crane
Crane-episodic-memory Aaron Sloman Last updated: 8 Mar 2009; 23 Oct 2011; 12 Jun 2013 (links added) Installed: 10 Jul 2007
NOTE:
This file ishttp://tinyurl.com/CogMisc/crane/
It is available as a stand alone PDF file<Seeing%5Fa%5FToy%5FCrane.pdf>A draft paper closely related to this is here:
http://www.cs.bham.ac.uk/research/projects/cosy/papers/#tr0801
TR-0801 Architectural and representational requirements for seeing processes and affordances. (PDF)For a partial overview of other online material here see
http://tinyurl.com/CogMisc/AREADME.html
Introduction
This is a sequel to two previous challenges for machine vision
- Seeing cups spoons and saucers
- Seeing impossible objects
- More impossible objects
- Seeing and thinking about loops in strings A presentation related to these topics can be found here:
http://www.cs.bham.ac.uk/research/projects/cogaff/talks/#compmod07
Architectural and representational requirements for seeing processes and affordances.Presented at BBSRC funded Workshop on
Closing the gap between neurophysiology and behaviour: A computational modelling approach
University of Birmingham, UK
May 31st-June 2nd 2007
The polyflap challenge is also relevanthttp://www.cs.bham.ac.uk/research/projects/cosy/papers/#tr0605
The Crane Challenge(s)
What follows is a collection of pictures of a toy crane made from plastic meccano (by Alison Sloman), viewed from various angles, along with pictures of some meccano parts.
The challenge is this: for every identifiable entity in any picture, small or large, find in other pictures
- Other entities of the same type (at various levels of abstraction)
- The very same entity (e.g. the same nut, or bolt, or wheel, or hook)
- Specify the 3-D position and pose of the object in the scene
- Describe its 3-D relations to other objects in the scene
- Find whether its shadow is visible in the scene, and if so identify the relevant portion of shadow. NOTE: merely identifying locations of parts _in the images_is of no interest in this challenge, though it may be an intermediate stage in processing, if found useful.
Related challenges:
- Wherever a crank handle is visible, work out what the consequences will be of turning it in either possible direction.
- Describe how the viewpoint is related to the objects in the scene, including the floor.
- Describe how the crane in the photographs differs from the crane depicted in the manual (see picture at end).
The Pictures
Picture {1} 
Picture {2} 
Picture {3} 
Picture {4} 
Picture {5} 
Picture {6} 
Picture {7} 
Picture {8} 
Picture {9} 
Picture {10} 
Picture {11} 
Picture {12} 
Picture {13} 
Picture {14} 
Picture {15} 
Picture {16} 
Picture {17} 
Picture {18} 
Picture {19} 
-----------
The diagram in the instruction manual.
(Not followed precisely.)
Meccano Online Web pages
http://www.internationalmeccanomen.org.uk
Home pagehttp://www.internationalmeccanomen.org.uk/EXHIB/Henley2008/
Pictures showing some amazing things built by Meccano enthusiasts in 2008.
See also
http://www.cs.bham.ac.uk/research/projects/cosy/photos/penrose-3d
http://www.cs.bham.ac.uk/research/projects/cosy/movies/sample-actions/
A multi-picture challenge for theories of vision, and machine vision systems
http://www.cs.bham.ac.uk/research/projects/cogaff/misc/multipic-challenge.pdf
Maintained byAaron Sloman
School of Computer Science
The University of Birmingham