Search papers, labs, and topics across Lattice.
K [1] and use the DIV, K test set together with 2,0002,000 additional images randomly sampled from COCO [21] and ImageNet [6] to evaluate the generalization performance. All images are resized to
1
0
3
Randomly throwing distortions at your watermarking model during training? Meta-FC shows meta-learning a better way, boosting robustness by up to 4.71% against combined distortions.
