How Deep is Your Art: An Experimental Study on the Limits of Artistic Understanding in a Single-Task, Single-Modality Neural Network

TL;DR

This study investigates the ability of a deep neural network to classify modern conceptual art and finds it relies solely on visual features like shape and color, ignoring contextual and intent-based properties.

Contribution

The paper experimentally validates that current DCNNs use only exhibited visual properties for art classification, highlighting limitations in understanding artwork meaning.

Findings

DCNN relies on exhibited properties like shape and color.

DCNN ignores non-exhibited properties such as context and artist intent.

Supports the hypothesis that current models lack multidimensional understanding.

Abstract

Computational modeling of artwork meaning is complex and difficult. This is because art interpretation is multidimensional and highly subjective. This paper experimentally investigated the degree to which a state-of-the-art Deep Convolutional Neural Network (DCNN), a popular Machine Learning approach, can correctly distinguish modern conceptual art work into the galleries devised by art curators. Two hypotheses were proposed to state that the DCNN model uses Exhibited Properties for classification, like shape and color, but not Non-Exhibited Properties, such as historical context and artist intention. The two hypotheses were experimentally validated using a methodology designed for this purpose. VGG-11 DCNN pre-trained on ImageNet dataset and discriminatively fine-tuned was trained on handcrafted datasets designed from real-world conceptual photography galleries. Experimental results supported the two hypotheses showing that the DCNN model ignores Non-Exhibited Properties and uses only Exhibited Properties for artwork classification. This work points to current DCNN limitations, which should be addressed by future DNN models.