This research paper evaluates various image denoising techniques to identify Rheumatic Heart Disease (RHD) symptoms, starting by addition of artificial noise to original images to display real-world conditions. Four filtering methods—Gaussian, Mean, Median, and a proposed filter—are applied to these noisy images. The Gaussian filter smoothens the images using a Gaussian kernel, whereas the suggested filter combines several methods for improved denoising, the mean filter average the values of nearby pixels, and the median filter substitutes the median of each pixel for its neighbors. The contrast of the images is then enhanced using Contrast Limited Adaptive Histogram Equalization (CLAHE). The effectiveness of each method is assessed using quantitative metrics such as MSE, AMBE, PSNR, Entropy, SSIM, and RMSE, which quantify noise removal and detail preservation. Results are visualized in graphs, highlighting the proposed filter’s superiority due to its balanced noise reduction and detail preservation. This paper underscores the importance of effective denoising techniques for accurate RHD symptom identification in medical images.