Towards Generative Art and Brain Computer Interfaces

This research work introduces a computational system for creating generative art using a Brain-Computer Interface (BCI) that portrays the user’s brain activity in a digital artwork. In this way, the user takes an active role in the creative process. The amalgam of each of the elements in the generated art pieces and their interrelation provide the necessary inputs that allow creating a generative art piece that is not only based on the brain activity of a person, but also fluctuates with it, so that, at first, none of the figures generated by the users are the same, since their mental states and/or brain activity vary constantly.
To create art pieces, circles were used because the low computational complexity required to draw them (Chen & Sundaram, 2005) allows including different characteristics into the generated art pieces, as shown in Figure 1.

Figure 1. User interaction.
It was decided to use the Neurosky’s Mindwave device to carry out this research because of the following reasons: The first is that Mindwave is a quality, non-invasive, and affordable BCI device. The second is that it uses dry electrodes and not saline. This is relevant since saline electrodes may cause interference in data acquisition. In users with abundant hair, some of the sensors may lose contact with the scalp and also interfere negatively in the user experience due to the wet sensation of the electrode. Finally, the third reason is that the main interest of this research work is to record the brain activity and to associate it with the user’s mental states, which is the reason why it is enough to obtain the information available in the frontal lobe from where Mindwave acquires brain data.
Multiple tests were carried out to determine which brain waves would provide more expressiveness and motion in order to generate unique art pieces. In this sense, it was decided to select Delta and Gamma waves since they are the pair that bring up more variability (Diaz Rincon, Reyes Vera, & Rodriguez C, 2019a). Delta waves were used to control the number of circles in the art pieces and gamma waves to control the radius of each circle.
This decision is also supported by the fact that these brain waves are the ones in a lower and higher frequency range respectively, as shown in the study conducted by (Abo-Zahhad, Ahmed, & Abbas, 2015). The above is reflected in the fact that having fewer circles that change rapidly in size provides greater contrast, variability, and motion in the art pieces.
Each circle moves in one of the two axes, following a straight line whose sinusoidal displacement and wavelength vary randomly. The device’s attention and relaxation values are used to choose, randomly, the opacity and color of each circle present in the art piece, which varies in a range that has the user’s attention level as lower bound and the relaxation level as upper bound.
As shown in the next video, the user’s brain waves allow creating art pieces with different morphological characteristics due to the number of variables that the algorithm gathers and that are expressed involuntarily. The geometric shapes that are formed contain a series of basic design concepts such as translation, superposition, and gradation of: shape, size, color, and scale.
http://tiny.cc/UserInteractionNS

Thanks to the presence of Artificial Intelligence at the feature extraction stage and in the acquisition and processing of encephalographic signals it was possible to make the most out of the user interaction knowing that the time for this is limited. In this regard, the developed algorithms were able to respond to the user’s mental states

The next video shows the projection of the generated figures over a sequence of pyramids in a spiral-shaped scale. This is done to offer a deep sensation while projecting the art pieces created by the interaction between the spectators and the BCI device. As an added value, the visual production is accompanied by a musical piece generated from the BCI data (Diaz Rincon, Reyes Vera, & Rodriguez C, 2019b) which complements the generated artwork providing a bimodal communication character.
http://tiny.cc/FrontalProjection

REFERENCES
Abo-Zahhad, M., Ahmed, S. M., & Abbas, S. N. (2015). A New EEG Acquisition Protocol for Biometric Identification Using Eye Blinking Signals. International Journal of Intelligent Systems and Applications. https://doi.org/10.5815/ijisa.2015.06.05
Chen, Y., & Sundaram, H. (2005). Estimating complexity of 2D shapes. In 2005 IEEE 7th Workshop on Multimedia Signal Processing. https://doi.org/10.1109/MMSP.2005.248668
Diaz Rincon, R. A., Reyes Vera, J. M., & Rodriguez C, P. J. (2019a). An approach to Generative Art from Brain Computer Interfaces. In Celestino Soddu (Ed.), XXII Generative Art Conference - GA2019 (pp. 332–343). Retrieved from http://www.generativeart.com/GA2019_web/50_RicardoDiaz_168x240.pdf
Diaz Rincon, R. A., Reyes Vera, J. M., & Rodriguez C, P. J. (2019b). Generando música a través de la Actividad Cerebral. Brazilian Journal of Development, 5(6), 5375–5388. https://doi.org/10.34117/bjdv5n6-077