Connecting Touch and Vision via Cross-Modal Prediction | IEEE Conference Publication | IEEE Xplore