SynthRender and IRIS: Open-Source Framework and Dataset for Bidirectional Sim-Real Transfer in Industrial Object Perception

TL;DR

This paper introduces SynthRender and IRIS, an open-source framework and dataset for improving object perception in industrial settings through bidirectional sim-to-real transfer, enabling more robust and data-efficient deep learning models.

Contribution

The work presents a novel integrated framework combining synthetic data generation with structured evaluation, including a new dataset IRIS, for systematic study of sim-to-real transfer in industrial perception.

Findings

Achieved 99.1% mAP@50 on a public robotics dataset

Achieved 98.3% mAP@50 on an automotive benchmark

Achieved 95.3% mAP@50 on the IRIS dataset

Abstract

Object perception is fundamental for tasks such as robotic material handling and quality inspection. However, modern supervised deep-learning models require large annotated datasets for robust automation under semi-uncontrolled conditions; a major barrier for widespread deployment with proprietary industrial parts. We address this through an integrated framework combining synthetic data generation and structured empirical evaluation for systematic investigation of bidirectional sim-to-real transfer. Our method integrates 2D-to-3D Reality-to-Simulation techniques for 3D asset creation from physical parts with programmatic Guided Domain Randomization (GDR) via SynthRender, an open-source synthetic image generation framework. Structured ablation studies across multiple benchmarks quantify the impact of individual rendering design choices, yielding practical guidelines for dataefficient synthetic training. To support evaluation under realistic industrial conditions, we introduce Industrial Real-Sim Imagery Set (IRIS), a 32-class dataset with diverse textures, intra-class variation, strong inter-class similarities, and 19,672 annotations, providing both CAD models and reconstructed meshes for bidirectional sim-to-real benchmarking. Across three industrial benchmarks, the proposed framework achieves highly competitive performance, reaching 99.1% mAP@50 on a public robotics dataset, 98.3% mAP@50 on an automotive benchmark, and 95.3% mAP@50 on IRIS.