SIMPL: Generating Synthetic Overhead Imagery to Address Zero-shot and Few-Shot Detection Problems
This addresses the problem of high data acquisition costs for researchers and practitioners in remote sensing and satellite imagery analysis, offering an incremental improvement over existing methods.
The paper tackles the challenge of acquiring training data for object detection in overhead imagery by introducing SIMPL, a method for generating synthetic training data, which improves detection in zero-shot and few-shot scenarios with specific performance gains.
Recently deep neural networks (DNNs) have achieved tremendous success for object detection in overhead (e.g., satellite) imagery. One ongoing challenge however is the acquisition of training data, due to high costs of obtaining satellite imagery and annotating objects in it. In this work we present a simple approach - termed Synthetic object IMPLantation (SIMPL) - to easily and rapidly generate large quantities of synthetic overhead training data for custom target objects. We demonstrate the effectiveness of using SIMPL synthetic imagery for training DNNs in zero-shot scenarios where no real imagery is available; and few-shot learning scenarios, where limited real-world imagery is available. We also conduct experiments to study the sensitivity of SIMPL's effectiveness to some key design parameters, providing users for insights when designing synthetic imagery for custom objects. We release a software implementation of our SIMPL approach so that others can build upon it, or use it for their own custom problems.