Generative abstract art has long been researched and enhanced by the advancement in computer software, graphics and Machine Learning. Some interesting sophisticated abstract paintings also have generated from the recent works, such as abstract Kandinsky-style  and Leonardo abstract paintings  using scalable and generic in generating pseudorandom images resembling the artist’s style.
This project has utilised the existing automated abstract painting works and to extend them to the next level. The program’s reads a music file and regenerate Kandinsky style abstract objects and shapes onto the canvas nested on the users screen in which the music will change their shape, size and colours according to the frequency.
Our Group PS2033 has built a Kandinsky styled program which displays a subtle but sleek UI allowing the user to insert a song of their choice when the program is initialised. Once the song has been entered the UI displays a number of chic coloured buttons which display age ranges as the increase in age choice will affect the speed of the shapes. Nested below such buttons lies two individual moods being happy and sad.
We decided to only utilise the two moods currently as we can include more colours rather than sharing them over more moods. Once the users designated song loads and the age range and mood is selected the program will initialise and randomly generate an amount of Kandinsky style objects/shapes which we have designed and implemented. As the song plays through the shapes will change size, colour and movement path as the frequency is being extracted and used to manipulate
 K. Zhang, J.H. Yu, Generation of Kandinsky Art. Leonardo, MIT Press, Vol.49, No.1, 2016, pp. 48-55.
 L. Xiong, K. Zhang, Generation of Miro’s Surrealism. In Proc. 9th Symp. on Visual Information Communication and Interaction (VINCI’2016), Dallas, USA, 24-26 September, 2016, ACM Press, pp. 130-137.
Developed by PS2033