Visual Understanding and Narration: A Deeper Understanding and Explanation of Visual Scenes
arXiv:1906.00038v2
Originality Synthesis-oriented
AI Analysis
This addresses the problem of enabling robots to interpret and narrate visual scenes for applications in autonomous navigation and human-robot interaction, but it appears incremental as it builds on existing visual question answering and captioning tasks.
The paper introduces the task of Visual Understanding and Narration, where a robot generates text to answer open-ended questions about images from its environment, aiming to provide deeper explanations of visual scenes.
We describe the task of Visual Understanding and Narration, in which a robot (or agent) generates text for the images that it collects when navigating its environment, by answering open-ended questions, such as 'what happens, or might have happened, here?'