Background <p>Digital image correlation (DIC) measured heterogeneous deformation fields promise enhanced validation of finite element (FE) models. In the existing literature, DIC-based FE model validation is commonly performed in a qualitative way based on visual appraisal of FE-DIC likeness of deformation fields. Moreover, DIC uncertainties are usually inadequately characterised.</p> Objective <p>This paper aims to take a step towards a quantitative approach by experimentally validating an enhanced DIC uncertainty quantification (UQ) procedure previously proposed by the authors.</p> Methods <p>Heterogeneous strain fields in a composite notched test sample were measured using DIC. DIC bias errors were accounted for through a numerical speckle deformation based digital twin (DT) of the DIC experiment. Heterogeneous DIC noise maps were used to formulate a fully-spatial model validation criterion. Careful test design choices were made to minimise the modelling errors originating from constitutive behaviour. Extensive investigation into the validity of the constitutive model was performed to justify the validation errors.</p> Results <p>The UQ approach, together with the application of DIC-measured boundary conditions in the FE model were found to be essential for proper validation. Decrease in the tangent modulus of the composite material caused by the accumulation of damage in the composite matrix was confirmed by cyclic tests. The FE constitutive model did not account for composite damage, resulting in validation errors consistent with these limitations.</p> Conclusions <p>The DIC-based FE model validation approach was proven to be effective when a specific protocol is followed.</p>