Accurate assessment of fetal head descent during labor is crucial for informed obstetric decision-making and optimizing delivery outcomes. Traditional ultrasound imaging methods for measuring key parameters, such as the angle of progression (AoP) and head-symphysis distance (HSD), often suffer from incomplete anatomical capture and susceptibility to noise, limiting their clinical utility. In this study, we propose a novel approach that leverages the Video Swin Transformer to integrate local, temporal, and global features from ultrasound videos, effectively capturing comprehensive anatomical information. Additionally, we incorporate wavelet transformers within a multitask learning framework to enhance noise mitigation and improve the robustness of AoP and HSD predictions. Experimental results on a clinical ultrasound dataset demonstrate that our method significantly outperforms existing models, achieving a reduction in mean absolute error of 18% for AoP and 22% for HSD measurements. This approach provides a robust and accurate solution for enhancing ultrasound assessments during labor, with the potential to improve clinical decision-making and patient outcomes in obstetric care.