Patch-Based Spatial Authorship Attribution in Human-Robot Collaborative Paintings

TL;DR

This paper introduces a patch-based method for attributing authorship in human-robot collaborative paintings, achieving high accuracy and providing a new approach for stylistic analysis in mixed-creation artworks.

Contribution

It presents a novel patch-based framework for spatial authorship attribution in collaborative paintings, demonstrating its effectiveness and potential for broader application.

Findings

88.8% patch-level accuracy in attribution

Outperforms texture-based baselines (68-84.7%)

Detects mixed authorship with 64% higher uncertainty in hybrid regions

Abstract

As agentic AI becomes increasingly involved in creative production, documenting authorship has become critical for artists, collectors, and legal contexts. We present a patch-based framework for spatial authorship attribution within human-robot collaborative painting practice, demonstrated through a forensic case study of one human artist and one robotic system across 15 abstract paintings. Using commodity flatbed scanners and leave-one-painting-out cross-validation, the approach achieves 88.8% patch-level accuracy (86.7% painting-level via majority vote), outperforming texture-based and pretrained-feature baselines (68.0%-84.7%). For collaborative artworks, where ground truth is inherently ambiguous, we use conditional Shannon entropy to quantify stylistic overlap; manually annotated hybrid regions exhibit 64% higher uncertainty than pure paintings (p=0.003), suggesting the model detects mixed authorship rather than classification failure. The trained model is specific to this human-robot pair but provides a methodological grounding for sample-efficient attribution in data-scarce human-AI creative workflows that, in the future, has the potential to extend authorship attribution to any human-robot collaborative painting.