SegEQA: Video Segmentation Based Visual Attention for Embodied Question Answering | IEEE Conference Publication | IEEE Xplore