PennSyn2Real: Training Object Recognition Models without Human Labeling

TL;DR

This paper introduces PennSyn2Real, a synthetic dataset and data generation framework that enables training object recognition models without human labeling, achieving competitive results and improving few-shot learning performance.

Contribution

The paper presents a scalable, artifact-free synthetic data generation method using chroma-keying and motion tracking, reducing the need for human-labeled data in training CNNs for object recognition.

Findings

Synthetic data achieves competitive detection and segmentation performance.

Bootstrapped synthetic data enhances few-shot learning accuracy.

Framework is easy to set up and applicable to various objects.

Abstract

Scalable training data generation is a critical problem in deep learning. We propose PennSyn2Real - a photo-realistic synthetic dataset consisting of more than 100,000 4K images of more than 20 types of micro aerial vehicles (MAVs). The dataset can be used to generate arbitrary numbers of training images for high-level computer vision tasks such as MAV detection and classification. Our data generation framework bootstraps chroma-keying, a mature cinematography technique with a motion tracking system, providing artifact-free and curated annotated images where object orientations and lighting are controlled. This framework is easy to set up and can be applied to a broad range of objects, reducing the gap between synthetic and real-world data. We show that synthetic data generated using this framework can be directly used to train CNN models for common object recognition tasks such as detection and segmentation. We demonstrate competitive performance in comparison with training using only real images. Furthermore, bootstrapping the generated synthetic data in few-shot learning can significantly improve the overall performance, reducing the number of required training data samples to achieve the desired accuracy.