*Grayscale Reconstruction

Pre-Processing > *Grayscale Reconstruction

Requires B/W Companion Image

Changes grayscale pixel values based on features in the Companion Image. Available modes are morphological reconstruction, impose minima (ensures that minima in the grayscale image only occur at feature pixels), or set pixels equal to the mean, median, or standard deviation of the pixels within each feature.

1. Mode

• Inpainting: Smoothly interpolates inward within each marked region from the pixel values on the outer boundary of the region. Works by computing the discrete Laplacian over the regions and solves the Dirichlet boundary value problem for each [1].

• Morphological Reconstruction: Performs morphological reconstruction using the defined markers [2]

• Impose Minima: Modifies the image so that it only has local minima wherever markers are defined

• Fill With Mean: Fills the pixels within the marker regions with the mean pixel value of that region

• Fill With Median: Fills the pixels within the marker regions with the median pixel value of that region

• Fill With Mode: Fills the pixels within the marker regions with the mode pixel value of that region

• Fill With StdDev: Fills the pixels within the marker regions with the standard deviation pixel value of that region

• Fill With Min: Fills the pixels within the marker regions with the minimum pixel value of that region

• Fill With Max: Fills the pixels within the marker regions with the maximum pixel value of that region

References

[1] Pérez, Patrick, Michel Gangnet, and Andrew Blake. “Poisson image editing.” In ACM Transactions on Graphics (TOG), vol. 22, no. 3, pp. 313-318. ACM, 2003. Harvard

[2] Vincent, L., “Morphological Grayscale Reconstruction in Image Analysis: Applications and Efficient Algorithms,” IEEE Transactions on Image Processing, Vol. 2, No. 2, April, 1993, pp. 176-201.