SIR-Net: Scene-Independent End-to-End Trainable Visual Relocalizer | IEEE Conference Publication | IEEE Xplore