Pharyngitis is a common inflammation of pharynx that often goes untreated in its early stages, potentially leading to complications if treatment is delayed. This study combines cutting-edge transformer approach with improved image processing techniques to offer precise detection of pharyngitis. Wavelet Transform (WT) based image preprocessing, is applied to the Pharyngitis dataset for decomposing an image into various frequency components. Bayesian based segmentation is utilized for segmenting the pre-processed images into regions by model uncertainty and unpredictability in pixel data. The Hilbert-Huang Transform (HHT), extract discriminative characteristics from pharyngitis throat images. A proposed Swin Transformer model, a cutting-edge vision transformer architecture that uses shifted windows for effective and hierarchical feature encoding. The Swin Transformer is to improve diagnostic accuracy by allowing the model to collect global structural information as well as fine-grained texture patterns. A Swin transformer framework is proposed to enhance the diagnosis of pharyngitis image. Using Python software, the proposed framework achieved higher accuracy and F1-score of 95%, precision and recall of 96%, is accomplished when compared to other techniques.