SynSpill: Improved Industrial Spill Detection With Synthetic Data

TL;DR

SynSpill introduces a synthetic data generation pipeline that enhances industrial spill detection by enabling effective fine-tuning of vision-language models and object detectors in data-scarce, safety-critical environments.

Contribution

The paper presents a scalable synthetic data framework that improves the performance of VLMs and detectors for industrial spill detection, addressing data scarcity and privacy concerns.

Findings

Synthetic data boosts VLM and detector performance in spill detection.

VLMs outperform detectors on unseen spill scenarios without synthetic data.

Combining synthetic data with lightweight fine-tuning yields comparable performance to real data.

Abstract

Large-scale Vision-Language Models (VLMs) have transformed general-purpose visual recognition through strong zero-shot capabilities. However, their performance degrades significantly in niche, safety-critical domains such as industrial spill detection, where hazardous events are rare, sensitive, and difficult to annotate. This scarcity -- driven by privacy concerns, data sensitivity, and the infrequency of real incidents -- renders conventional fine-tuning of detectors infeasible for most industrial settings. We address this challenge by introducing a scalable framework centered on a high-quality synthetic data generation pipeline. We demonstrate that this synthetic corpus enables effective Parameter-Efficient Fine-Tuning (PEFT) of VLMs and substantially boosts the performance of state-of-the-art object detectors such as YOLO and DETR. Notably, in the absence of synthetic data (SynSpill dataset), VLMs still generalize better to unseen spill scenarios than these detectors. When SynSpill is used, both VLMs and detectors achieve marked improvements, with their performance becoming comparable. Our results underscore that high-fidelity synthetic data is a powerful means to bridge the domain gap in safety-critical applications. The combination of synthetic generation and lightweight adaptation offers a cost-effective, scalable pathway for deploying vision systems in industrial environments where real data is scarce/impractical to obtain. Project Page: https://synspill.vercel.app