Pulmonary fibrosis (PF) is a chronic lung disorder in which the lung tissue becomes permanently scarred of early detection and accurate segmentation plays an important role in proper diagnosis and planning of the best treatment. However most existing approaches often provide only quantitative predictions without generating detailed binary, color-coded, and overlay segmentation outputs that can assist clinicians in visualizing fibrosis-affected regions. To address this issue, a novel Deep Learning (DL)-based AUD PFS model for PF segmentation is suggested, which uses CT images from the OSIC Pulmonary Fibrosis Progression Dataset. The hybrid Median with Weiner (MedWe) filter is used to improve the CT image and to remove the noise in the images. ResNet50 is used to extract the feature from the CT image for reliable segmentation and Attention UNet is employed for segmenting the CT images of PF as Binary segmentation, color-codded segmentation and Overlay segmentation. The proposed AUD PFS is evaluated using the following metrics: accuracy, precision, recall, specificity, F1-score, Dice index, and Jaccard index. From the experimental analysis, the proposed AUD PFS model attains an overall accuracy of 99.02% and F1-score of 97.23% for efficient PF segmentation. Moreover, the proposed AUD PFS model improves the overall accuracy by 1.53%, 0.02%, 13.85%, 9.11% and 5.17% better than ResNet18-based U-Net, nnU-Net, HCR-DL model, CNN, and AE-IPF respectively.