Expansive Soil Desiccation Crack Segmentation using DeepLabv3+

DeepLabv3+ with different backbones were used to perform expansive soil recognition and segmentation. The zip included the code and sample image datasets used to train, validate, and test the DeepLabv3+ with 4 different backbones (MobileNetV2, ResNet-18, ResNet-50, and Xception). The code was implemented using MATLAB R2023b. See README.txt for more information.

本数据集采用搭载不同骨干网络（backbone）的DeepLabv3+模型开展膨胀土识别与分割任务。该压缩包包含用于训练、验证及测试搭载4种不同骨干网络（MobileNetV2、ResNet-18、ResNet-50与Xception）的DeepLabv3+模型的代码与示例图像数据集。本次所用代码基于MATLAB R2023b实现。更多详细信息请参阅README.txt文件。