Action Recognition Utilizing YGAR Dataset

TL;DR

This paper introduces a new 3D action data simulation engine and datasets to address the scarcity of high-quality action videos, enabling more flexible research in action recognition.

Contribution

A novel 3D data simulation engine and datasets that facilitate advanced action recognition research and model training.

Findings

Demonstrated the engine's application to image classification and action recognition

Generated datasets that support complex action recognition tasks

Validated the datasets with common recognition models

Abstract

The scarcity of high quality actions video data is a bottleneck in the research and application of action recognition. Although significant effort has been made in this area, there still exist gaps in the range of available data types a more flexible and comprehensive data set could help bridge. In this paper, we present a new 3D actions data simulation engine and generate 3 sets of sample data to demonstrate its current functionalities. With the new data generation process, we demonstrate its applications to image classifications, action recognitions and potential to evolve into a system that would allow the exploration of much more complex action recognition tasks. In order to show off these capabilities, we also train and test a list of commonly used models for image recognition to demonstrate the potential applications and capabilities of the data sets and their generation process.