The rising prevalence of colon cancer necessitates early detection through colonoscopy. Deep learning segmentation technology has emerged as a reliable tool for detecting lesions, with convolutional neural networks (CNNs) leading advancements in medical image processing. However, manual polyp segmentation during colonoscopic examinations is time-consuming, highlighting the need for automated approaches. This study introduces a novel parallel branching structure to address limitations and improve contextual information and low-level detail handling. The structure extracts complementary features, enriching image analysis. Additionally, a quadratic complete volume integral branch enhances segmentation performance. We present our innovative model and methodology for polyp segmentation, advancing automated detection in colonoscopy. Experimental results demonstrate the effectiveness and robustness of our approach. By integrating deep learning techniques and leveraging our parallel branching structure, our method achieves superior segmentation accuracy for efficient and accurate polyp detection.