An Improved and Efficient Approach for Enhancing the Precision of Diagnostic CT Images

In non-contrast computed tomography (CT) systems, acquired low contrast CT scans are a common issue that reduces clear visibility of the picture and resists the way of evoking its essential facts. As a result, an improved CT image enhancement scheme for non-contrast CT scans is proposed, which produces pleasing output CT images suitable for medical diagnosis. The entire procedure is divided into two major modules: contrast enhancement and DWT-based fusion with noise reduction. To begin, an exposure-centered contrast-restricted bi-histogram equalization technique with adaptive threshold is proposed, yielding a global contrast-enhanced output CT image while preserving critical image information. Simultaneously, to emphasize important minor information in the CT picture, local enhancement with global intensity is operated to the original CT image. Additionally, to acquire an appropriate contrast-enhanced CT image, a discrete wavelet transform (DWT)-based fusion scheme is used to combine the global and local-enhanced CT images, as well as a denoised filter to exclude extraneous noise in the CT image. Experiments on a wide range of CT images were performed to assess the proposed method's success, both qualitative and quantitative. Substantial quantitative analysis demonstrates that the proposed approach outperforms state-of-the-art enhancement techniques in terms of Discrete entropy, contrast index, signal to noise ratio, and similarity measure. Contrast is enhanced although visibility and visual clarity are maintained with the proposed algorithm. As a result, the proposed procedure produces a higher-quality CT picture suitable for analysis and diagnosis.