New benchmark MDTD-ArtIR evaluates art image restoration models

A new benchmark, MDTD-ArtIR, has been developed to evaluate image editing and restoration models specifically for art image restoration under complex degradations like cracks and stains. The benchmark utilizes a new, open-source dataset of alpha texture masks called MDTD-Art. Experiments show that general image editing models, when guided by structured prompts, outperform specialized restoration architectures in preserving details and structural consistency, highlighting the importance of recoverable semantic information and prompt controllability. AI

IMPACT This benchmark could drive improvements in AI models for preserving and restoring cultural heritage assets.

RANK_REASON The cluster describes a new academic paper introducing a benchmark and dataset for a specific AI task. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• Gotit.pub
• Hugging Face
• MDTD-Art
• MDTD-ArtIR
• Mridula Vijendran
• ScienceCast

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →