Perceptually fine detail extraction from a vector field

In this paper, a perceptually weighted quadratic optimization problem is formulated for the extraction of fine details from a vector field. To achieve better edge-preserving performance, a weighted sum square error (SSE) is adopted in the proposed optimization problem. The weighted SSE is used as data term in the proposed quadratic optimization problem to refine the extraction strength according to local intensity variance. Experimental results show that the perceptually weighted optimization problem can minimize sharp edge information in the extracted details.