Magnetic Resonance Imaging (MRI) is a widely used medical imaging technique. It has many different applications that ranges from manual diagnosis process by medical experts to automated disease predictions. Sparsity of MRI helped researchers in generating rapid images from highly under-sampled signals. Moreover, it is being used to reduce input feature set for various learning algorithms. Traditionally implemented techniques use different transform domains for sparsity of MRI. This paper presents a novel technique for improving sparsity of MR images using Slant transform and compares the results with traditionally used sparse domains.