Seismic simulation of paleokarst systems and deep learning for characterizing paleokarst features in 3D seismic images

To train our deep convolutional neural network for paleokarst delineation in 3D seismic images, we automatically generate 120 pairs of synthetic training datasets including 3D seismic images and the corresponding label images with the ground truth of the paleokarst systems simulated in the seismic images. 1) The "seismic.zip" contains 120 3D seismic images, each image is with the dimension of 256X256X256; 2) The "karst.zip" contains 120 3D label images of the karsts. Each label image is with the same dimension of 256X256X256. The values in a label image are set with ones in the karst areas while zeros elsewhere, which is why the compressed label images in the karst.zip is much smaller than the seismic images compressed in the seismic.zip为训练用于三维地震图像中古岩溶（paleokarst）圈定的深度卷积神经网络，我们自动生成了120组合成训练数据集，包含三维地震图像及其对应的标注图像，标注图像中带有模拟于地震图像内的古岩溶系统的真值标注（ground truth）。 1) "seismic.zip" 包含120张三维地震图像，单张图像尺寸为256×256×256； 2) "karst.zip" 包含120张岩溶三维标注图像，每张标注图像的尺寸与前述地震图像一致，均为256×256×256。标注图像中，岩溶区域的像素值设为1，其余区域设为0，这正是"karst.zip"的压缩文件体积远小于"seismic.zip"压缩文件的原因。