In this paper a principal neighborhood dictionary nonlocal means method isproposed.As the computational power increases, data-driven descriptions of structure are becoming increasingly important in image processing. Traditionally, many models are used in applications such as denoising and segmentation have been based on the assumption of piecewise smoothness. Unfortunately, these models yields limited performance thus motivated for data driven strategies. One data-driven strategy is to use image neighborhoods for representing local structure and these are rich enough to capture the local structures of real images, but do not impose an explicit model. This representation has been used as a basis for image denoising and segmentation. But the drawback is it gives high computational cost. The motivation of our work is to reduce the computational complexity and higher the accuracy by using nonlocal means image denoising algorithm. This paper will present in-depth analysis of nonlocal means image denoising algorithm that uses principal component analysis to achieve a higher accuracy while reducing computational load. Image neighborhood vectors are projected onto a lower dimensional subspace using PCA. The dimensionality of this subspace is chosen automatically using parallel analysis. Consequently, neighborhood similarity weights for denoising are computed using distances in this subspace rather than the full space. The resulting algorithm is referred to as principal neighborhood dictionary nonlocal means. By implementing the algorithm we will investigateprincipal neighborhood dictionary nonlocal meansmethod’sandNonlocal Means method’s accuracy with respect to the image neighborhood and window sizes. Finally, we will present a quantitative and qualitative comparison of both methods.