The LMN model ranked 1st place on MovieQA Video+Subtt-based Answering Challenge 2017 (The Joint Video and Language Understanding Workshop, ICCV 2017).
- The flowchart of Layered Memory Network (LMN).
- The framework of Dynamic Subtitle Memory module with update mechanism.
python mqa_video+subtitle+update+question.py
Bo Wang, Youjiang Xu, Yahong Han, Richang Hong. "Movie Question Answering: Remembering the Textual Cues for Layered Visual Contents." AAAI, 2018. [Paper]
@inproceedings{Wang2018,
author = {Bo Wang and Youjiang Xu and Yahong Han and Richang Hong},
title = {Movie Question Answering: Remembering the Textual Cues for Layered Visual Contents},
booktitle = {AAAI},
year = {2018},
}