Exploring the Effectiveness of Dataset Synthesis: An application of Apple Detection in Orchards

TL;DR

This study evaluates the use of Stable Diffusion 2.1-base for generating synthetic apple orchard images to train object detection models, showing promising results with minimal performance gap compared to real data.

Contribution

It demonstrates the feasibility of using synthetic data generated by Stable Diffusion for training effective apple detection models, reducing reliance on extensive real-world datasets.

Findings

Synthetic data yields nearly comparable detection accuracy to real data.

The model performs well except under heavy shading conditions.

Synthetic dataset creation is a promising alternative for data collection.

Abstract

Deep object detection models have achieved notable successes in recent years, but one major obstacle remains: the requirement for a large amount of training data. Obtaining such data is a tedious process and is mainly time consuming, leading to the exploration of new research avenues like synthetic data generation techniques. In this study, we explore the usability of Stable Diffusion 2.1-base for generating synthetic datasets of apple trees for object detection and compare it to a baseline model trained on real-world data. After creating a dataset of realistic apple trees with prompt engineering and utilizing a previously trained Stable Diffusion model, the custom dataset was annotated and evaluated by training a YOLOv5m object detection model to predict apples in a real-world apple detection dataset. YOLOv5m was chosen for its rapid inference time and minimal hardware demands. Results demonstrate that the model trained on generated data is slightly underperforming compared to a baseline model trained on real-world images when evaluated on a set of real-world images. However, these findings remain highly promising, as the average precision difference is only 0.09 and 0.06, respectively. Qualitative results indicate that the model can accurately predict the location of apples, except in cases of heavy shading. These findings illustrate the potential of synthetic data generation techniques as a viable alternative to the collection of extensive training data for object detection models.