We address the limitations of Deep learning models for 3Dgeometry segmentation by using Conditional Random fields (CRF). Weshow that CRFs can take advantage of the neighbouring structure of pointclouds to assist the learning of the Deep Learning models (DL). Our hybridPN-CRF model is able to learn more optimal weights by taking advantageof equal-segmentation assignments to neighbouring points. As a result,it increases the robustness in the model specially for segmentation taskswhere correctly detecting the boundaries between segmentations is veryimportant.