3D vision is a good example to choose to discuss representation. It feels like, and is usually thought of, as a process of reconstruction. It seems as if our brain has reconstructed a version of the outside world that we can explore with our imagination and that underlies our perception of the scene.
The terms 'head-centred', 'body-centred' and 'world-centred' refer to putative representations in parietal cortex and hippocampus. They are supposed to have an origin and 3 axes. The task in these pages is to determine whether ideas like this can be re-cast in terms of the restricted tools that are proposed here (e.g in words or symbols). It is useful to do this in a series of steps:
- rotation of the eye/camera about its optic centre
- small translation of the eye/camera in different directions but considering only a small patch of the image
- small translation of the eye/camera in different directions but considering only the changing angles between objects
- larger translations
The argument is that none of these require a 3-dimensional representation. The hope is that all of these levels can be united within a single, non-Cartesian framework based on the restricted tools referred to above.