iNEDI - Improved New Edge-Directed Interpolation Algorithm

iNEDI is an acronym for "improved New Edge-Directed Interpolation." This function implements significant enhancements to the original NEDI technique described in the paper "New Edge-Directed Interpolation" by X. Li et al., published in IEEE Transactions on Image Processing, Vol. 10, No. 10, October 2001, pp. 1521-1527. The algorithm improves upon the original NEDI method by optimizing edge detection and directional interpolation processes. Key implementation features include enhanced covariance matrix calculations for more accurate edge orientation estimation and improved interpolation weights determination. The code typically involves sophisticated gradient computation, directional filtering, and adaptive interpolation kernels that better preserve edge sharpness while reducing artifacts. iNEDI provides a more efficient and accurate image interpolation approach suitable for various image processing applications such as image scaling, super-resolution, and digital zoom. The implementation often includes optimized matrix operations and edge-aware interpolation routines that maintain structural integrity while upsampling images.