Newanimalfarm1984

DCT-DIGITAL-watermark-

The digital watermarking of DCT algorithm based on Arnold scrambling, which utilize Matlab as a programming tool, it is programmed to achieve and developed a related research. The watermarking image is reconstructed by Arnold, and then the DCT is carried out the watermark image and the host image. The embedding of the watermark is carried out, and the host image is embedded in the watermark to obtain the carrier image. The watermark is extracted by the carrier image and the original host image, then the DCT coefficients of the carrier image are obtained and the DCT coefficients of the host image are obtained. Then, we compare the watermark image to be extracted on the carrier image and the host image. After the DCT inverse transform and Arnold anti-scrambling are carried out, the extracted watermark image. Utilizing the telescopic, rotate, add the Gaussian noise and other attacks to verify the watermark of the anti-attack. The simulation results demonstrate that the algorithm can achieve the requirement of robustness and invisibility. the algorithm is safe that the anti-attack performance of the watermark is superior to the general DCT algorithm.

deeplab_v3

使用deeplab_v3模型对遥感图像进行分割

Image-Semantic-segmentation

models

Models and examples built with TensorFlow

remote-sensing-datasets

satellite_seg

CCF BCDI2017 卫星影像的AI分类与识别

Semantic-Quantization

Tensorflow implementation for Training and Inference for Integer-based Semantic Segmentation

semantic_segmentation_contest

遥感图像稀疏表征与智能分析竞赛之语义分割

tensorflow-deeplab-v3-plus

DeepLabv3+ built in TensorFlow

Newanimalfarm1984/DCT-DIGITAL-watermark-

The digital watermarking of DCT algorithm based on Arnold scrambling, which utilize Matlab as a programming tool, it is programmed to achieve and developed a related research. The watermarking image is reconstructed by Arnold, and then the DCT is carried out the watermark image and the host image. The embedding of the watermark is carried out, and the host image is embedded in the watermark to obtain the carrier image. The watermark is extracted by the carrier image and the original host image, then the DCT coefficients of the carrier image are obtained and the DCT coefficients of the host image are obtained. Then, we compare the watermark image to be extracted on the carrier image and the host image. After the DCT inverse transform and Arnold anti-scrambling are carried out, the extracted watermark image. Utilizing the telescopic, rotate, add the Gaussian noise and other attacks to verify the watermark of the anti-attack. The simulation results demonstrate that the algorithm can achieve the requirement of robustness and invisibility. the algorithm is safe that the anti-attack performance of the watermark is superior to the general DCT algorithm.

Newanimalfarm1984/deeplab_v3

使用deeplab_v3模型对遥感图像进行分割

Newanimalfarm1984/Image-Semantic-segmentation

Newanimalfarm1984/models

Models and examples built with TensorFlow

Newanimalfarm1984/remote-sensing-datasets

Newanimalfarm1984/satellite_seg

CCF BCDI2017 卫星影像的AI分类与识别

Newanimalfarm1984/Semantic-Quantization

Tensorflow implementation for Training and Inference for Integer-based Semantic Segmentation

Newanimalfarm1984/semantic_segmentation_contest

遥感图像稀疏表征与智能分析竞赛之语义分割

Newanimalfarm1984/tensorflow-deeplab-v3-plus

DeepLabv3+ built in TensorFlow