Using ShapeNet and Pascal Voc dataset we estimation rotation and translation(camera extrinsic parameters). Using CNN as pose regressor and the project is implemented in pytorch.I used L1 loss and Adam optimizer with a learning rate of 1e-4.