The precise identification and delineation of polyps in medical images, particularly those derived from colonoscopy and endoscopy, hold the utmost importance in the field of gastrointestinal disease diagnosis and treatment. However, manual image annotation by qualified gastroenterologists is laborious and vulnerable to errors. To combat this issue, we present an advanced deep learning framework for automated polyp segmentation. Our framework enhances the PolypSegNet architecture by incorporating a squeeze and excitation module into the convolutional layer block. This integration improves feature discrimination and reduces model complexity. Additionally, we augment the deep reconstruction module with a feature pyramid network to further amplify polyp segmentation performance. Training our model on the Kvasir-SEG dataset, which comprises polyp images and segmentation masks by expert medical practitioners, allows for accurate localization, precise boundary demarcation, and differentiation from non-polyp structures. The proposed deep learning framework represents a significant advancement in gastrointestinal disease diagnosis and treatment. With its accurate localization, precise boundary delineation, differentiation from non-polyp structures, variability handling, real-time processing, generalization, and integration with clinical workflows, our model revolutionizes polyp detection and segmentation. Medical practitioners can improve patient care and outcomes by leveraging this powerful tool.