Before the invention of photography, botanical illustration was the only way to visually record the many species of plants on the planet. Today scientific documents and books are full of detailed photographs to discover the fascinating forms of flora but before these technological discoveries, artists and illustrators were the ones responsible for spreading the beauty of botany in the world. Botanical illustration was a job that required great artistic skills, attention to the smallest details and tremendous horticultural culture. 24th Japan Media Arts Festival’s Jury Selection Winner, ‘Artificial Botany’ by fuse* is an ongoing project which explores the latent expressive capacity of botanical illustrations through the use of machine learning algorithms.
Artificial Botany’s system draws from public domain archive images of illustrations by the greatest artists of the genre, including Maria Sibylla Merian, Pierre-Joseph Redouté, Anne Pratt, Mariann North, and Ernst Haeckel; plus, fuse* have found open-source images of plants that could be tested freely on the Biodiversity Heritage Library, with a large collection of botanical illustration. These illustrations have become the learning material for two GANs, which through a training phase are able to recreate new artificial images and descriptions with morphological elements almost identical to the images of inspiration, but with details and features as if they were generated by human painters. The machine in this sense re-elaborates the content by creating a new language, capturing the information and artistic qualities of human and nature. The caption underlying each artwork is generated by the exploitation of another neural network algorithm called “image to text translation”: while commonly it’s used to classify images, here it has been tested by asking it to recognize other artificial-generated images frame by frame.
Artists and researchers have long explored the expressive abilities of generative computational systems. One interesting aspect of these systems lies in their intrinsic emerging characteristics, subtending a marked homeostatic and self-organizational tendency. The input data is therefore configured as a variable parameter, generating variable and locally diversified behaviours. The information learned by the software within this multidimensional, latent space serves the network to be able to select particular stylistic features from the input data set, learning precisely to recognize what it receives and possibly be able to reproduce it. Not exactly a copy of the original in the present case, but a version aesthetically reinterpreted by the algorithm, only “stylistically” like the original. This process highlights the artistic potential of a new and totally synthetic aesthetic, and the dataset with which was decided to teach the software to recognize certain stylistic sequences.
Awards: Jury Selections of the 24th Japan Media Arts Festival, Digital Design Awards 2020
Art Direction: Mattia Carretti, Luca Camellini
Concept: Mattia Carretti, Luca Camellini, Samuel Pietri
Supervisione Software: Luca Camellini
Software Artists: Luca Camellini, Samuel Pietri
Sound Design: Riccardo Bazzoni
Hardware Engineering: Matteo Mestucci
Support to Concept Writing: Saverio Macrì