DSPNet: Dual-vision Scene Perception for Robust 3D Question Answering

Jingzhou Luo, Yang Liu, Weixing Chen +4

3D Question Answering (3D QA) requires the model to comprehensively understand its situated 3D scene described by the text, then reason about its surrounding environment and answer…