A significant number of people are susceptible to pneumonia, a potentially fatal respiratory illness, especially in places with high pollution, dense populations, unhygienic environments and inadequate medical facilities. Pericardial effusion, a condition where fluid accumulates in the chest and impairs breathing, is usually the result of pneumonia. Pneumonia ought to be diagnosed promptly and accurately in order to treated effectively and increase survival rates. Pneumonia is often manually detected by specialists, but this method is laborious and prone to human error, making it ineffective for processing a large number of images. The proposed framework encompasses four stages: image preprocessing, segmentation, feature extraction and classifier. In this study an effective image preprocessing approach, Contrast Limited Adaptive Histogram Equalization (CLAHE) is implemented for reducing the noise amplification. K-means clustering is employed for dividing the noise-free image into numerous regions. Inception V_3 is used to classify the features that were extracted misinterpretation and improve patient outcomes, while Wavelet Transform (WT) is employed to extract the best characteristics of the segmented image for precise and effective diagnosis. An efficient implementation of Python is used to assess the methodology’s overall performance. The data collected indicate that the suggested classifier results with a maximum accuracy of 95.67% compared to the existing classifiers.