Non-Local Means is an image denoising algorithm based on patch similarity. It compares a reference patch with the neighboring patches to find similar patches. Such similar patches participate in the weighted averaging process. Most of the computational time for Non-LocalMeans is consumed to measure patch similarity. In this thesis, we have proposed an improvement where the image patches are projected into a global feature space. Then we have performed a statistical t-test to reduce the dimensionality of this feature space.Denoising is achieved based on this reduced feature space and the proposed modification exploits an improvement in terms of denoising performance and computational time.