Abstract. Among numerous techniques of information concealment, Spread spectrum watermarking (SSW) has proved to yield improved results when robustness against attack is at a premium. SSW hides information which is called watermark by spreading its spectrum and then adding them to a host image as a watermarked image. Spreading spectrum is done by a pseudo-noise (PN) sequence. But in standard SSW approaches, receiver must use a separate channel to achieve PN sequence used at the transmitter for detecting hidden information. The unique PN sequence must have a low cross correlation and thereby can be duplicated easily by hostile attackers. In this paper a novel approach based on the unconventional Random Encoding Spread Spectrum is proposed for recovering the spreading sequence of watermark signal without any information from the transmitter. It is contributed a higher secure feature by using of the time-varying random encoded spread spectrum.

Convolutional neural networks (CNN) based object detection usually assumes that training and test data have the same distribution, which, however, does not always hold in real-world applications. In autonomous vehicles, the driving scene (target domain) consists of unconstrained road environments which cannot all possibly be observed in training data (source domain) and this will lead to a sharp drop in the accuracy of the detector. In this paper, we propose a domain adaptation framework based on pseudo-labels to solve the domain shift. First, the pseudo-labels of the target domain images are generated by the baseline detector (BD) and optimized by our data optimization module to correct the errors. Then, the hard samples in a single image are labeled based on the optimization results of pseudolabels. The adaptive sampling module is approached to sample target domain data according to the number of hard samples per image to select more effective data. Finally, a modi ed knowledge distillation loss is applied in the retraining module, and we investigate two ways of assigning soft-labels to the training examples from the target domain to retrain the detector. We evaluate the average precision of our approach in various source/target domain pairs and demonstrate that the framework improves over 10% average precision of BD on multiple domain adaptation scenarios on the Cityscapes, KITTI, and Apollo datasets.

No abstract

No abstract