Filtering And Enhancement Of Lung Computerized Tomography Images For The Diagnosis Of Covid-19 And Lung Cancer

Abstract

Recently, the rapid spreading of coronavirus (Covid-19) has created a pandemic situation all around the world. Early diagnosis is the only way to control the spread of coronavirus. Another major threat to the human community is the increasing number of lung cancer patients. In this complex scenario, lung computerized tomography (CT) images can be effectively used for early identification of Covid-19 patients and lung cancer patients as well. These CT images are processed using Computer Aided Diagnosis (CAD) techniques by the usage of suitable algorithms. To improve the accuracy of this diagnosis, effective restoration and enhancement techniques are essential. In this paper, a novel algorithms is proposed for CT image restoration and enhancement. Specifically, a new algorithm called 2D Improved Anisotropic Diffusion Bilateral Filter (2D IADBF) is proposed for the effective noise removal from the CT lung images. The proposed restoration algorithm is compared with other algorithms like 2D Median Filter, 2D Log Filter and 2D Frequency Domain Wavelet Filter. This comparison is performed using filtration metrics like peak signal to noise ratio (PSNR) in dB, mean square error (MSE)and processing time. For the enhancement of the restored images, a new algorithm called 2D Edge Preservation Efficient Histogram Improvement (2D EPEHI) algorithm is proposed. The efficacy of this algorithm is proved by comparison with other algorithms like Contrast Limited Adaptive Histogram Equalization (CLAHE), 2D Adaptive Mean Adjustment and Image Coherence improvement algorithms. The enhancement algorithms were evaluated using metrics like Structural Similarity Index (SSIM) and Absolute Mean Brightness Error (AMBE). Performance analysis clearly show that the proposed restoration and enhancement algorithms produces best results compared to other traditional algorithms. Thus, these algorithms can be effectively used in the clinical pathway for the early diagnosis of Covid-19 and lung cancer.