what ever we learn visually up to 10 years of age it is permanently stored in the storage device in area no: 18 and 19 of brain the capacity of this area depends on extent of learning. images stored from 11 to 25 years of age may be long lasting but may not be permanent because memory neurons development is very slow and with advancing age it further decreases. when we see the same images in future these are compared with the stored images and recognized. all this is processed in area no: 17 of brain. if area 18 and 19 are damaged one can see but can not recognize, he is practically blind but do not accept blindness . this occurs in Anton's syndrome. the images are stored in the form of center and surround ? where center may be off and surround on or vice versa, like 0 and 1 language of computer

Here's a recent article on "bag of visual words".

I wasn't aware before that this term existed.

http://www.technologyreview.com/view/537786/is-this-the-first-computational-imagination/

http://arxiv.org/abs/1505.05190

Note that the Technologyreview pages might

show a nag screen which can be avoided

by not having JavaScript enabled.

Similar approaches, papers:

http://www.technologyreview.com/view/532231/random-image-experiment-reveals-the-building-blocks-of-human-imagination/

http://www.newswise.com/articles/view/516254/

http://www.eurekalert.org/pub_releases/2011-02/jhu-jft021011.php

http://www.eurekalert.org/pub_releases/2007-07/gumc-ftf073107.php

http://www.nature.com/nature/journal/v442/n7102/abs/nature04951.html

http://arxiv.org/abs/1410.4627

https://idw-online.de/de/news247546

http://phys.org/news/2010-11-phantom-images-flexible-network-brain.html

http://www.cell.com/current-biology/abstract/S0960-9822%2811%2900937-7

Regards,

Joachim

Dear Nadine as far as I know much of the debate on visual imagery has shifted to the A.I. and robotics research field, albeit many theoretical insights still come from traditional competing theories (i.e. Kosslyn-like vs. Pylyshyn-like). But if you look for some models it may be worthwhile having a look. Here it is a small sample:

1.Aleksander, I., e Morton, H.B. (2007a). Depictive architectures for synthetic phenomenology. In A. Chella, e R. Manzotti (a c. di), Artificial consciousness. Imprint Acaddemic: Exter, pp. 30-45.

2. Aleksander, I., e Morton, H.B. (2007b). Phenomenology and digital neural architectures. Neural Networks, 20, pp. 932-937.

3. Di Nuovo, A., De La Cruz, V.M., e Marocco, D. (2013). Special issue on artificial mental imagery in cognitive systems and robotics. Adaptive Behavior, 21, 4, 217-221.

4. Marques, G.H., e Holland, O. (2009). Architectures for functional imagination. Neurocomputing, 72, pp. 743-759.

From the psychological point of view, there is still a sort of phenomenological third way between picture-like and propositional accounts; the empirical issues are usually in papers written in italian and are not so current as maybe you wish  but you may look at:

5. Cornoldi, Beni, Giusberti, Massironi, Memory and Imagery, A Visual Trace is not a Mental image. In Martin et al., Theories of Memory, Taylor & Francis, 1998.