TVQA+: Spatio-Temporal Grounding for Video Question Answering
This work addresses the need for interpretable AI systems that can understand videos by jointly retrieving moments and detecting objects, though it is incremental as it builds on existing datasets and tasks.
The authors tackled the problem of spatio-temporal video question answering by augmenting the TVQA dataset with 310.8K bounding boxes and proposing the STAGE framework, which achieved effective performance in grounding evidence to answer questions about videos.
We present the task of Spatio-Temporal Video Question Answering, which requires intelligent systems to simultaneously retrieve relevant moments and detect referenced visual concepts (people and objects) to answer natural language questions about videos. We first augment the TVQA dataset with 310.8K bounding boxes, linking depicted objects to visual concepts in questions and answers. We name this augmented version as TVQA+. We then propose Spatio-Temporal Answerer with Grounded Evidence (STAGE), a unified framework that grounds evidence in both spatial and temporal domains to answer questions about videos. Comprehensive experiments and analyses demonstrate the effectiveness of our framework and how the rich annotations in our TVQA+ dataset can contribute to the question answering task. Moreover, by performing this joint task, our model is able to produce insightful and interpretable spatio-temporal attention visualizations. Dataset and code are publicly available at: http: //tvqa.cs.unc.edu, https://github.com/jayleicn/TVQAplus