IRHMapNet Radargram and Mask Patches Dataset.

# README: IRHMapNet Radargram and Mask Patches Dataset ## Dataset Overview This dataset contains radargram patches and corresponding mask patches used for training and evaluating the **IRHMapNet** model. The dataset is designed for segmentation of internal reflection horizons (IRHs) from radio-echo sounding data. The data is organized into two directories: radargram patches (`grams_patches`) and mask patches (`masks_patches`), with each patch having dimensions of 512x512 pixels. ### Contents - **grams_patches/**: Contains 600 `.csv` files representing radargram patches. Each file is a 512x512 matrix corresponding to a small section of the radargram image.- **masks_patches/**: Contains 600 `.csv` files representing the ground-truth mask patches for segmentation. Each file is a 512x512 binary mask, where `1` indicates the presence of an internal reflection horizon (IRH), and `0` represents background or ice. ## Data Format - The files in both directories are named consistently, with matching pairs of radargram and mask patches. - Example: `grams_patches/patch_001.csv` corresponds to `masks_patches/patch_001.csv`.- Each `.csv` file is a comma-separated values (CSV) file containing 512 rows and 512 columns. ## Directory Structure ```DATA_IRHMapNet/├── grams_patches/ # Radargram patches│ ├── patch_001.csv│ ├── patch_002.csv│ └── ... (600 patches)└── masks_patches/ # Mask patches (Ground truth) ├── patch_001.csv ├── patch_002.csv └── ... (600 patches)``` ## Usage Instructions 1. **Loading the data**: Each `.csv` file can be loaded using standard CSV reading functions in Python, such as `numpy.loadtxt()` or `pandas.read_csv()`. Example in Python using `numpy`: ```python import numpy as np radargram_patch = np.loadtxt('grams_patches/patch_001.csv', delimiter=',') mask_patch = np.loadtxt('masks_patches/patch_001.csv', delimiter=',') ``` 2. **Model training**: These patches are designed for input into a U-Net or similar convolutional neural network architectures for pixel-wise classification tasks. The radargram patches serve as input, and the mask patches provide the ground-truth labels for training. ## License This dataset is made available under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. You are free to: Share — copy and redistribute the material in any medium or format. Adapt — remix, transform, and build upon the material for any purpose, even commercially. You must give appropriate credit by citing the following publication: **Citation**: Moqadam, H., et al. (2024). Going deeper with deep learning: Automatically tracing internal reflection horizons in ice sheets. *Journal of Geophysical Research: Machine Learning and Computation*. DOI: [insert DOI] ## Contact For questions or further information, please contact Hameed Moqadam at [[e-mail]]. Data Curator: Hameed Moqadam Annotator: Hameed Moqadam Data Manager: Hameed Moqadam