Answers are grounded in this paper and cite their sources.
Deep learning-based displacement measurement methods provide an end-to-end solution for automatic and real-time displacement measurement. However, existing deep learning-based displacement measurement methods typically rely on single-scale speckle images, which severely compromise their accuracy and adaptability in measuring displacement. This paper proposes a multi-scale displacement network (MSDN) for automatic measurement of displacement fields using camera systems. The innovative architecture of the proposed multi-scale neural network model merges the displacements identified at various individual scales using a novel fusion module, thereby overcoming the inherent limitations of single-scale models. A tailored training strategy for the multi-scale model is then provided. Subsequently, extensive experimental analyses were conducted to validate its performance and adaptability. The results validate that the proposed method has high precision in measuring displacement fields across a spectrum of deformations, from minor to substantial.
No takes yet. Share an insight, caveat, or question.
Xiong et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: