Image of the Forest
“(...) all intelligence is an abstraction of the present context. Therefore producing an abstraction is the essence of intelligence. But that abstraction is only a snapshot of the organism; it is not the organism itself. All models are wrong, because we build them to perform actions that are not feasible using the original.”
~The Oxford Handbook of Ethics of AI
Moreover, human visual system is wired to organise visual input in terms of increasingly complex image features: edges, shapes, and structures.
The bottom-up visual pathway starts at the lowest level of the image — edges. From there, subsequent regions of the visual cortex 'build up' the image, constructing shapes and structures. The images we see can, therefore, be understood in terms of these basic features of a system of "wholeness"— in which, certain geometric properties and structures have a universal ability to evoke specific emotional states.
To understand how the “images of the forest” are perceived in the visual cortex and how it impacts emotional states, we’ve created a dataset that represents images segmented into their features and organised on a scale of increasing complexity. Original photographs represent the highest level of complexity—structures.
By utilizing the discrete Fourier transform to analyze spectra of the analog forest photographs, we seek the understanding of how different complexity levels within an image can trigger changes in neural response.