StoryTeller: Training-Free Narrative Grounding for Long-Form Audio Description
2026-07-13 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors created StoryTeller, a system that helps describe movies for blind and low-vision viewers by remembering important characters and story details across scenes. Unlike other models that only look at short clips, StoryTeller keeps track of the whole story to make descriptions more connected and meaningful. It works without needing scripts or extra data and uses public movie info carefully checked against the video. The authors also made a way to test if these descriptions really capture the story by asking questions about them. Their tests show StoryTeller makes better, clearer audio descriptions than previous methods.
audio descriptionvideo-language modelsnarrative memorysemantic filteringstory contextblind and low-vision accessibilityquestion-answering benchmarkmovie metadatalong-form video analysis
Authors
Seung Hyun Hahm, Minh T. Dinh, SouYoung Jin
Abstract
Long-form audio description (AD) requires more than describing visible actions: it must preserve characters, events, relationships, and story context across scenes so that blind and low-vision (BLV) audiences can follow a film. Modern video-language models (VLMs) are effective on short clips, but they often treat each moment independently, producing descriptions that miss who characters are, why events matter, and how the current scene connects to earlier narrative context. We propose StoryTeller, a training-free framework for story-aware long-form AD. Instead of relying only on local visual cues, StoryTeller maintains a verified narrative memory that carries forward story-relevant information across scenes, enabling later descriptions to remain coherent, grounded, and contextually informative. Given only raw video and a movie title, StoryTeller can optionally retrieve public movie metadata to resolve names and story context, while accepting only facts that are supported by the video through semantic filtering and VLM verification. The method requires no subtitles, scripts, AD transcripts, aligned captions, character banks, precomputed face identities, or task-specific fine-tuning. To evaluate whether generated AD preserves narrative information, we introduce StoryAD-QA, a question-answering benchmark that tests whether a language model can answer story-context questions using only the generated descriptions. Experiments on standard AD benchmarks and diverse long-form videos show that StoryTeller consistently improves narrative coherence, factual grounding, and story comprehension over strong baselines in automatic, QA-based, and human evaluations.