Moment-preserving sharpening—A new approach to digital picture deblurring

Ling-Hwei Chen, Wen-Hsiang Tsai
1988 Computer Vision Graphics and Image Processing  
A new approach to sharpening blurred pictures using the moment-preserving principle is proposed. The new gray value assigned to each pixel in the desired picture is computed deterministically in such a way that the moments of the n x n neighborhood of the pixel are preserved. The selection of the neighborhood size n depends on the blurring degree of the picture. The sharpening operation can also be applied iteratively to an input picture to get better sharpening effect. Experimental results
more » ... that the approach is indeed effective for picture deblurring. Boolean combinations of VLSI masks are important in VLSI design. We describe an optimal algorithm for two layers and rectangles having the same orientation in the plane. The algorithm runs in time 0( n log n + p) using space O(n), where n is the number of rectangles and p is the number of edges in the output. Note that this algorithm avoids spending computation time on edges that do not contribute to the output, unlike its counterpart for arbitrary polygons. Moreover, the contour of the Boolean combination is computed in the form of contour cycles rather than single contour edges. Also, we extend the algorithm to more than two layers and to orthogonal polygons having the same orientation. The time bound remains the same for any fixed number of layers; it is 0( k2k n log n + p), for k 2 2 layers when k is an input parameter. Q 1988 Academic PWS. IIIC.
doi:10.1016/0734-189x(88)90114-4 fatcat:4jukmqn3wfdq3i7qthwne2bose