A Hybrid Enhancement Framework for Chest X-Ray Images Using CLAHE, Haar Wavelet, and Denoising Filters: A Quality Metrics Analysis

Chest radiographic imaging is one of the more frequently used diagnostic techniques in the diagnostic investigation of diseases. This study suggests a hybrid image enhancement technique that combines the contrast-limited adaptive Histogram Equalization (CLAHE) with Haar wavelet transformation, along without and with different noise-removing filters to enhance the chest X-ray images. To achieve better local contrast while preserving edges, the framework involves an adaptive clip limit in the CLAHE algorithm and an adaptive K-factor that is used to enhance Haar waveform detail. Besides, bilateral noise reduction, Wiener, nonlocal mean (NLM) filters, and cascade noise filters were explored to remove the image noise while maintaining the significant structural information. Typical chest X-ray images, both in normal and pneumonia conditions, were used to evaluate the proposed enhancement methods. Objective measures such as peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), entropy, contrast, sharpness, and noise contrast have been used to quantify the image quality. Experimental results indicate that adaptive modification of the CLAHE clip limit and Haar wave K index greatly enhances the quality of images in comparison to conventional enhancement. Moreover, applying the right filters for noise reduction can preserve the structure and reduce noise amplification without sacrificing diagnostic information and benefit to all quality measures, especially with the cascade bilateral filter with a Wiener filter and bilateral, nonlocal filter alone. The results of the comparative analysis show that the proposed hybrid framework always outperforms the others in terms of the different quality measures, yielding improved visualization of the lung structures.