Digital Image Correlation (DIC) has been widely used as a non-contact deformation measurement technique. Nevertheless, its accuracy is greatly affected by the speckle pattern on the specimen. To systematically evaluate how speckle deformability affects the precision of DIC algorithms. In this study, a test dataset of 2D speckle patterns with various prescribed deformation fields was numerically generated, containing two categories of speckles, i.e., the deformable and the non-deformable (rigid) ones. This dataset was used to evaluate the performance of inverse compositional Gauss-Newton (ICGN)-based DIC algorithms with two types of shape function (first-order and second-order), in the different scenarios of the deformation field. The results showed that imaging noise had a significant influence on the DIC algorithm. The first-order shape function (ICGN-1) performed better when tracking the simple linear deformation field. While the second-order shape function (ICGN-2) was proved to perform better on non-linear deformations. Moreover, the deformability of the speckle was found to have an obvious impact on the performance of the DIC algorithm. ICGN-2 could effectively reduce so-called speckle rigidity induced (SRI) error. Conclusively, ICGN-2 should be chosen as priority, because of its feasibility on non-linear deformation fields and speckle rigidity. While in the linear deformation scenarios, ICGN-1 was still a robust and efficient method.
