In this study, we explore the integration of eye-gaze (EG) information obtained via eye-tracking sensors to enhance medical image segmentation. EG is additional information on an image, and to incorporate EG information in medical image segmentation, we apply Gaussian blurring masked by the collected attention maps generated by eye tracking. The variance of distribution on each pixel is adjusted in a certain way by EG information. After applying Gaussian blurring, classic models including UNet, FCN, and other models were trained and compared. The results indicate that incorporating EG information in addition to preprocessing data yields superior performance on certain metrics, demonstrating notable advantages in accurately identifying isolated polyps that grow on the surface of a human tissue. This innovative approach highlights the potential of combining sensor-derived data with advanced image processing techniques to improve medical diagnostics.