Application of computer image processing technology in old artistic design restoration

Area of Science:

• Digital image processing
• Computer vision
• Art conservation

Background:

• Artworks degrade over time, losing original textures and details.
• Digital technologies offer new methods for art restoration and preservation.
• Reviving legacy art requires advanced image processing techniques.

Purpose of the Study:

• To introduce a novel Textural Restoration Technique (TRT) for degraded artworks.
• To leverage Deep Feature Processing (DFP) for accurate texture and gradient restoration.
• To enhance the digital revival of historical art pieces.

Main Methods:

• Analyzing tampered images to extract available textures and features.
• Identifying missing gradients based on sequential texture patterns and gradient distribution.
• Employing recurrent learning to verify gradient substitutions and ensure texture evenness.
• Classifying texture patterns using high and low accuracy features between regions of interest (ROIs).

Main Results:

• Successful identification and filling of missing gradients within ROIs.
• Accurate sketching of textural edges by combining available and missing features.
• Achieving maximum restoration through a two-stage recurrent learning process.
• Quantifying restoration accuracy using a defined restoration ratio based on filled edges.

Conclusions:

• The proposed TRT with DFP effectively restores degraded textures in artworks.
• Recurrent learning ensures accurate gradient substitution for seamless texture completion.
• This technique significantly aids in preserving and reviving artistic legacies through digital means.