MultiDimensional Analysis of An Oil Painting Center of
- Slides: 19
Multi-Dimensional Analysis of An Oil Painting Center of Imaging Science Rochester Institute of Technology Pano Spiliotis May 7, 1999 11/5/98 Pano Spiliotis
Overview • • • Background Procedure Results Evaluation of data Conclusion 11/5/98 Pano Spiliotis
Background • Multi-dimensional segmentation. – Using image capturing protocols and multidimensional segmentation, artifacts that are present in works of art will be easily identified. • New approach in finding artifacts in oil paintings. 11/5/98 Pano Spiliotis
Overview • • • Background Procedure Results Evaluation of data Conclusion 11/5/98 Pano Spiliotis
Camera Setup • • Monochrome CCD camera Ultraviolet Black Light (Blue light) Tungsten Light Source Infrared Filter Important that images are in registration. 11/5/98 Pano Spiliotis
Illumination • Three types – Ultraviolet – Visible – Infrared 11/5/98 Pano Spiliotis
Ultraviolet Illumination • Ultraviolet Illumination – New Varnish – Over-painting 11/5/98 Pano Spiliotis
Visible Illumination • Topographical Inspection – Canvas Repair – Buckling of Canvas – Over-painting 11/5/98 Pano Spiliotis
Infrared Illumination • Infrared filter is presented to the system. – Under-drawing – Areas that are not present in the UV and VIS. 11/5/98 Pano Spiliotis
Flat Fielding • Procedure taking a non-uniformly illuminated image and resulting in an evenly illuminated image. Flat-fielded image non-uniformly illuminated image 11/5/98 Pano Spiliotis
Image Segmentation • • Looking for common areas. Segmentation of a common area. Trial and error. Save common area as an image and pixel value. 11/5/98 Pano Spiliotis
Overview • • • Background Procedure Results Evaluation of data Conclusion 11/5/98 Pano Spiliotis
Results Visible image 11/5/98 Infrared image Pano Spiliotis Ultraviolet image
2 -D Segmentation Program Visible Image Infrared Image 11/5/98 Pano Spiliotis
Pseudo Coloring Visible Image Infrared Image 11/5/98 Pano Spiliotis
After Segmentation Visible Image Infrared 11/5/98 Pano Spiliotis
Overview • • • Background Procedure Results Evaluation of data Conclusion 11/5/98 Pano Spiliotis
Evaluation of results • Areas that have been segmented reveal artifacts. • John Sutton, Associate Conservator from The West Lake Conservators, was called in to prove these findings. • The halo is not the original halo painted. 11/5/98 Pano Spiliotis
Conclusion • Research proved to be successful. • 2 -D Segmentation provides art conservator’s with a tool to validate their discoveries. 11/5/98 Pano Spiliotis
