ABSTRACT A pelvic fracture may lead to multiple complications such as hemorrhage with a high mortality rate. Rapid diagnosis and assessment of injury levels are critical for survival and reduction of morbidity. AP plain x-rays are widely used in the pelvis's rapid evaluation, identification, and primary fixation. In this paper, we retrained the Mask R-CNN model using x-ray images of the pelvis to detect maxillary fractures and measure the angle between the hip bones, which helps classify the trauma mechanism. The method yielded satisfactory experimental results. Keywords: Multiple sequence alignment, Pairwise sequence identity, Sequence selection treatment, Transmembrane protein, Transmembrane segment