An Action Is Worth Multiple Words: Handling Ambiguity in Action Recognition

TL;DR

This paper tackles the ambiguity in action recognition by developing methods to train multi-label models from single positive labels, using human-annotated pseudo-labels based on feature similarity, and introduces new benchmarks for evaluation.

Contribution

The authors propose two novel approaches leveraging human-annotated pseudo-labels for training multi-label action recognition models from single positive labels, and create new benchmarks for evaluation.

Findings

Outperform existing methods on new benchmarks

Effective use of human-annotated pseudo-labels

Improved multi-label action recognition accuracy

Abstract

Precisely naming the action depicted in a video can be a challenging and oftentimes ambiguous task. In contrast to object instances represented as nouns (e.g. dog, cat, chair, etc.), in the case of actions, human annotators typically lack a consensus as to what constitutes a specific action (e.g. jogging versus running). In practice, a given video can contain multiple valid positive annotations for the same action. As a result, video datasets often contain significant levels of label noise and overlap between the atomic action classes. In this work, we address the challenge of training multi-label action recognition models from only single positive training labels. We propose two approaches that are based on generating pseudo training examples sampled from similar instances within the train set. Unlike other approaches that use model-derived pseudo-labels, our pseudo-labels come from human annotations and are selected based on feature similarity. To validate our approaches, we create a new evaluation benchmark by manually annotating a subset of EPIC-Kitchens-100's validation set with multiple verb labels. We present results on this new test set along with additional results on a new version of HMDB-51, called Confusing-HMDB-102, where we outperform existing methods in both cases. Data and code are available at https://github.com/kiyoon/verb_ambiguity