Customized HouseExpo dataset

This dataset was prepared for our research on global map completion in robotics. It is designed for tasks involving sequential and incremental map data collection during robot exploration.The dataset is based on the HouseExpo environment and was generated using Gazebo and ROS to simulate exploration sequences with enhanced frontier detection.The dataset includes:3,384 images for training376 images for validation940 images for testingAll images are in PNG format, converted from the original PGM format to ensure ease of use.Additional implementation details, code, and usage instructions can be found on the accompanying GitHub repository:👉 https://github.com/f0120050/houseexpo_based_custom_dataset_with_gazebo