Exploring Modulated Detection Transformer as a Tool for Action Recognition in Videos

TL;DR

This paper investigates the application of Modulated Detection Transformer (MDETR), a multi-modal model, to action detection in videos without prior training, demonstrating its potential for generalization to new tasks.

Contribution

The study explores using MDETR for action detection in videos without training, highlighting its ability to generalize to tasks beyond its original design.

Findings

MDETR can be applied to action detection without training.

The model shows potential for generalizing to new downstream tasks.

Quantitative results obtained on the Atomic Visual Actions dataset.

Abstract

During recent years transformers architectures have been growing in popularity. Modulated Detection Transformer (MDETR) is an end-to-end multi-modal understanding model that performs tasks such as phase grounding, referring expression comprehension, referring expression segmentation, and visual question answering. One remarkable aspect of the model is the capacity to infer over classes that it was not previously trained for. In this work we explore the use of MDETR in a new task, action detection, without any previous training. We obtain quantitative results using the Atomic Visual Actions dataset. Although the model does not report the best performance in the task, we believe that it is an interesting finding. We show that it is possible to use a multi-modal model to tackle a task that it was not designed for. Finally, we believe that this line of research may lead into the generalization of MDETR in additional downstream tasks.