Laplacian deformation with Label Smoothing Method for 3D Animation Expression Synthesis

3D facial animations is essential for augmentation and mixed digital media and evaluated aesthetic components for potent story-telling media. It is difficult task for developing the real 3D animations which correctly replicates the individual’s speech and specific talking styles to animate the speech-driven face. The previous approaches showed effective performance, but still struggled to attain the lifelike mimicry. In this manuscript, developed the Laplacian deformation with Label Smoothing (LS) technique for synthesis of 3D facial animations. The LS is incorporated in the process of Laplacian deformation, that smoothens the solution through reducing the displacement. The dataset used for the research are BIWI and VOCASET datasets which includes the 3D facial images for animation synthesis. The Laplacian deformation with LS technique attained high accuracy of 98.54% and less error rate of 1.33 and attained high accuracy of 85.47% and less error rate of 16.57. The Laplacian deformation with LS technique showed the better performance while compared with existing techniques.