Paper Title
Revitalization of Computed Tomography Image of Lungs

Abstract
Computed Tomography (CT) plays a vital role in the early detection and diagnosis of lung diseases. However, due to dose limiting acquisition protocols and limitations from hardware constraints, these diagnoses are confounded by low contrast and noisy images with lower resolution. To overcome these limitations, this work proposes a hybrid image enhancement pipeline that leverages both classical image processing methods and deep learning methods. The hybrid pipeline involves Contrast Limited Adaptive Histogram Equalization (CLAHE) for constrast enhancement, wavelet transform and Non-Local Means (NLM) for reduction in noise, and Laplacian filtering for enhancement of edges. Finally, an Efficient Sub-Pixel Convolutional Neural Network (ESPCN) was trained to achieve super- resolution reconstruction flow- resolution input images. The experimental results showed improvements with respect to Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), and visual perusal than baseline methods. The enhanced images allowed for improved feature extraction for support with subsequent tasks such as segmentation and classification while supporting adequate early diagnosis of lung disease with lung images. Keywords - Computed Tomography (CT), Lung Disease, Image Enhancement, CLAHE, Wavelet Denoising, Non-Local Means (NLM), Laplacian Filtering, ESPCN, PSNR, SSIM, Deep Learning, Medical Image Processing.