Alzheimer's disease (AD) is a neurodegenerative disorder that primarily affects the elderly, leading to significant memory and cognitive impairments. As there is currently no cure for AD, early and accurate detection is critical. MRI scans are a valuable supportive tool in assessing structural brain changes associated with AD, but they are often compromised by artefacts that can degrade scan quality and negatively impact diagnostic accuracy. To address this, pre-processing of MRI scans is essential to enhance their quality and improve classification performance. This paper proposes a novel four-step pre-processing technique: RSTS. The ADNI database has been used to evaluate the proposed technique. The proposed method is primarily assessed based on classification accuracy, and the results demonstrated its effectiveness in supporting the diagnosis of Alzheimer's disease.