# M-VAD Names: a Dataset for Video Captioning with Naming

## TL;DR

This paper introduces M-VAD Names, an enhanced dataset with character identities for movie captioning, and demonstrates its use in training models to accurately name characters in video captions.

## Contribution

The paper presents an improved version of the M-VAD dataset with character annotations and a semi-automatic annotation process for video captioning.

## Key findings

- The dataset contains 63k visual tracks and 34k textual mentions with character identities.
- Multimodal architectures can effectively replace generic tags with proper character names.
- Evaluation on external videos shows the approach's generalization capability.

## Abstract

Current movie captioning architectures are not capable of mentioning characters with their proper name, replacing them with a generic "someone" tag. The lack of movie description datasets with characters' visual annotations surely plays a relevant role in this shortage. Recently, we proposed to extend the M-VAD dataset by introducing such information. In this paper, we present an improved version of the dataset, namely M-VAD Names, and its semi-automatic annotation procedure. The resulting dataset contains 63k visual tracks and 34k textual mentions, all associated with character identities. To showcase the features of the dataset and quantify the complexity of the naming task, we investigate multimodal architectures to replace the "someone" tags with proper character names in existing video captions. The evaluation is further extended by testing this application on videos outside of the M-VAD Names dataset.