Video Stream Retrieval of Unseen Queries using Semantic Memory

TL;DR

This paper introduces a novel approach for live video stream retrieval using semantic memory, enabling real-time search of unseen queries by leveraging pre-trained classifiers and adaptive memory techniques.

Contribution

It proposes a no-example, semantic relatedness-based retrieval method with memory pooling and welling to handle shifting video content in live streams.

Findings

Effective retrieval on large-scale video datasets

Successful handling of real-time, unseen queries

Improved performance over traditional methods

Abstract

Retrieval of live, user-broadcast video streams is an under-addressed and increasingly relevant challenge. The on-line nature of the problem requires temporal evaluation and the unforeseeable scope of potential queries motivates an approach which can accommodate arbitrary search queries. To account for the breadth of possible queries, we adopt a no-example approach to query retrieval, which uses a query's semantic relatedness to pre-trained concept classifiers. To adapt to shifting video content, we propose memory pooling and memory welling methods that favor recent information over long past content. We identify two stream retrieval tasks, instantaneous retrieval at any particular time and continuous retrieval over a prolonged duration, and propose means for evaluating them. Three large scale video datasets are adapted to the challenge of stream retrieval. We report results for our search methods on the new stream retrieval tasks, as well as demonstrate their efficacy in a traditional, non-streaming video task.