The Block Matching and 3D Filtering (BM3D) algorithm is widely used as one of the suitable methods for image denoising, due to its ability to effectively remove noise without affecting the original image. In BM3D techniques, block matching is achieved by square blocks and Wiener filter for filtering, respectively. In use of the BM3D for rectangular images, the images need to be resized to fit the square blocks’ mechanism. This paper has adopted a flexible approach for image denoising with rectangular blocks for addressing the problem mentioned above. In this paper, the computationally demanding collaborative Wiener filtering is replaced by an efficient total variation (TV) filter. The TV filter requires little tuning of the parameters compared to the Wiener filter. Compared to the inverse transform computation for square blocks, computing the rectangular blocks’ pseudo-inverse increases for large size of the images’ computational costs. The proposed method produces both flexibility and speed for image denoising.