CogStream: Context-guided Streaming Video Question Answering

TL;DR

This paper introduces CogStream, a new streaming video reasoning task that emphasizes relevance-based context selection, supported by a new dataset and baseline model, to improve efficiency and accuracy in real-world scenarios.

Contribution

It proposes the CogStream task, creates a densely annotated dataset, and develops CogReasoner, a model that improves streaming video question answering by selecting relevant context.

Findings

CogReasoner effectively handles streaming video reasoning tasks.

The dataset enables detailed evaluation of context relevance.

Relevance-guided context selection improves model performance.

Abstract

Despite advancements in Video Large Language Models (Vid-LLMs) improving multimodal understanding, challenges persist in streaming video reasoning due to its reliance on contextual information. Existing paradigms feed all available historical contextual information into Vid-LLMs, resulting in a significant computational burden for visual data processing. Furthermore, the inclusion of irrelevant context distracts models from key details. This paper introduces a challenging task called Context-guided Streaming Video Reasoning (CogStream), which simulates real-world streaming video scenarios, requiring models to identify the most relevant historical contextual information to deduce answers for questions about the current stream. To support CogStream, we present a densely annotated dataset featuring extensive and hierarchical question-answer pairs, generated by a semi-automatic pipeline. Additionally, we present CogReasoner as a baseline model. It effectively tackles this task by leveraging visual stream compression and historical dialogue retrieval. Extensive experiments prove the effectiveness of this method.