The Painting Fool Teaching Interface

Aus de_evolutionary_art_org
Version vom 14. November 2015, 13:50 Uhr von Gubachelier (Diskussion | Beiträge)

(Unterschied) ← Nächstältere Version | Aktuelle Version (Unterschied) | Nächstjüngere Version → (Unterschied)
Wechseln zu: Navigation, Suche


Simon Colton: The Painting Fool Teaching Interface. In: Computational Creativity 2010 ICCC 2010. p 289.



The Painting Fool is software that we hope will be taken seriously as a cre- ative painter in its own right – one day. As we are not trained artists, a valid criticism is that we are not equipped to train the software. For this reason, we have developed a teaching interface to The Painting Fool, which enables anyone – including artists and designers – to train the software to generate and paint novel scenes, according to a scheme they specify. In order to specify the nature and rendering of the scene, users must give details on some, or all, of seven screens, some of which employ AI techniques to make the specification process simpler. The screens provide the following functionalities: (i) Images: enables the usage of context free design grammars to generate images. (ii) Annotations: enables the annotation of digital images, via the labelling of user-defined regions. (iii) Segmentations: enables the user to specify the parameters for image segmenta- tion schemes, whereby images are turned into paint regions. (iv) Items: enables the user to hand-draw items for usage in the scenes, and to specify how each exemplar item can be varied for the generation of alternatives. (v) Collections: enables the user to specify a constraint satisfaction problem (CSP) via the ma- nipulation of rectangles. The CSP is abduced from the rectangle shapes, colours and placements, and when solved (either by a constraint solver or evolutionary process), generates new scenes of rectangles, satisfying the user constraints. (vi) Scenes: enables the specification of layers of images, items, segmentations and collections, in addition to substitution schemes. (v) Pictures: enables the spec- ification of rendering schemes for the layers in scenes. In the demonstration, I will describe the process of training the software via each of the seven screens. I will use two running example picture schemes, namely the PresidENTS series and the Fish Fingers series, exemplars of which are portrayed in figure 1.

Extended Abstract


author = {Simon Colton},
title = {The Painting Fool Teaching Interface},
editor = {Dan Ventura, Alison Pease, Rafael P ́erez y P ́erez, Graeme Ritchie and Tony Veale},
booktitle = {Proceedings of the First International Conference on Computational Creativity},
series = {ICCC2010},
year = {2010},
month = {January},
location = {Lisbon, Portugal},
pages = {289},
url = {, },
publisher = {International Association for Computational Creativity},
keywords = {computational, creativity},

Used References



Full Text

intern file

Sonstige Links