Computer vision training dataset generation for robotic environments using Gaussian splatting

TL;DR

This paper presents a pipeline using 3D Gaussian Splatting and game engine physics to generate large, realistic, and labeled datasets for robotic vision, reducing manual annotation and domain gap issues.

Contribution

It introduces a novel dataset generation method combining 3D Gaussian Splatting with advanced rendering and physics simulation for realistic, automatically labeled data in robotic environments.

Findings

Hybrid training with real and synthetic data improves model performance.

The pipeline produces high-quality, automatically labeled datasets.

Realism enhancements lead to better detection accuracy.

Abstract

This paper introduces a novel pipeline for generating large-scale, highly realistic, and automatically labeled datasets for computer vision tasks in robotic environments. Our approach addresses the critical challenges of the domain gap between synthetic and real-world imagery and the time-consuming bottleneck of manual annotation. We leverage 3D Gaussian Splatting (3DGS) to create photorealistic representations of the operational environment and objects. These assets are then used in a game engine where physics simulations create natural arrangements. A novel, two-pass rendering technique combines the realism of splats with a shadow map generated from proxy meshes. This map is then algorithmically composited with the image to add both physically plausible shadows and subtle highlights, significantly enhancing realism. Pixel-perfect segmentation masks are generated automatically and formatted for direct use with object detection models like YOLO. Our experiments show that a hybrid training strategy, combining a small set of real images with a large volume of our synthetic data, yields the best detection and segmentation performance, confirming this as an optimal strategy for efficiently achieving robust and accurate models.