|
Briefing
|
- AI Detecting Art Forgeries – Researchers from Rutgers University and The Atelier for Restoration and Research of Paintings in the Netherlands used deep recurrent neural network (RNN) and trained machine-learning algorithm to spot art forgeries through stroke features of various artists
- Technique Combination – RNN learned important stroke features (i.e. artist pushing, weight of line) of almost 300 artists with 80,000 individual strokes, enabling trained machine-learning algorithm to match specific features in artwork to more accurately detect forgeries that humans cannot
- Expected Outcome – Inspire art historians and researchers, steeped in centuries of tradition, to embrace more AI applications in art
- Technology Limitation – Current technique can only be used when lines are obvious and proves to be unhelpful where brushstrokes are invisible, with researchers planning to test methods on Impressionist works and other 19th-century art where brush strokes are clearer to validate results
|
