Traditional 2D face recognition encounters difficulties with pose variations, while 3D face recognition necessitates the acquisition of 3D image data. To surmount these challenges, we propose a model that leverages 3D reconstruction to achieve pose-invariant face verification using 2D images as input. Initially, we employ 2D feature matching techniques to extract pose- and expression-robust features from facial images. Subsequently, we integrate 3D reconstruction models to derive depth-based features from 2D images, thereby enhancing the verification process. Our approach combines the outcomes of 2D feature matching with those of 3D reconstructed models of both reference and probe images, effectively mitigating pose-related challenges without the necessity for 3D data acquisition. We meticulously curated a diverse dataset comprising subjects of varying ages, and genders, with and without spectacles, and under various lighting conditions. This dataset encompasses 250 samples from 25 subjects. Validation of the proposed model involved standard dataset such as the FEI Face dataset alongside our proprietary dataset. We achieved an overall verification accuracy of 99.63%, and an overall identification accuracy of 92.77%.