New Study Finds Over 96% of Computer Vision (CV) Teams Already Using Synthetic Data for Training and Testing of Visual Machine Learning Models