In this paper, we propose a novel DNN watermarking method that utilizes a\nlearnable image transformation method with a secret key. The proposed method\nembeds a watermark pattern in a model by using learnable transformed images and\nallows us to remotely verify the ownership of the model. As a result, it is\npiracy-resistant, so the original watermark cannot be overwritten by a pirated\nwatermark, and adding a new watermark decreases the model accuracy unlike most\nof the existing DNN watermarking methods. In addition, it does not require a\nspecial pre-defined training set or trigger set. We empirically evaluated the\nproposed method on the CIFAR-10 dataset. The results show that it was resilient\nagainst fine-tuning and pruning attacks while maintaining a high\nwatermark-detection accuracy.\n