Model implementation, models with trained weights, and sample inversion results of "Implicit Neural Representation to Improve the Resolution of 3D Electrical Resistivity"

Detailed Description of Each File1. INR.zipContents: Raw data used in the Implicit Neural Representation method (e.g., input coordinates, observed data) and the corresponding model outputs after training (e.g., reconstructed resistivity fields, predicted values). Format: ZIP archive; after extraction, typical files include .npy, .mat, .txt, .csv, or custom formats. Usage: Reproduce INR model predictions or serve as input for further analysis. 2. Numerical.zipContents: Raw data and model outputs from the Res2Dinv and Res3Dinv methods (e.g., synthetic/field ERT data and inverted resistivity distributions). Format: ZIP archive; internal file structure follows the standard input/output format of the respective methods. Usage: Compare with INR results to validate the proposed method. 3. Implicit Neural Representation.ipynbType: Jupyter Notebook Function: Implements the complete training pipeline for the Implicit Neural Representation method, including data loading, network definition, loss function, training loop, and model saving. Outputs: Training logs, loss curves, intermediate results, and final model weights (likely saved inside INR.zip or a separate folder). Dependencies: See "Environment Requirements" below. 4. Forward Modeling ERT (update).ipynbType: Jupyter Notebook Function: Performs ERT forward modeling to compute potential/apparent resistivity responses given a resistivity model. Includes updated parameter settings, solver configurations, and result visualization. Outputs: Forward response data, mesh information, synthetic datasets (which may be used as inputs in INR.zip or Numerical.zip). Applicability: Serves as a common data generation source for both INR and traditional inversion methods. 5. Figures and Analysis of ERT-3D-INR paper.ipynbType: Jupyter Notebook Function: Generates all figures in the main text of the *ERT-3D-INR* paper (e.g., comparison plots, error maps, cross-sections, convergence curves) and performs statistical analysis of the results. Inputs: Reads output files from INR.zip, Numerical.zip, and the forward modeling notebook. Outputs: Exportable .png, .pdf, or vector graphics files identical to those shown in the paper. Usage InstructionsDownload and extractUnzip INR.zip and Numerical.zip into the current directory (or keep them zipped if the scripts support auto-reading). Set up the environment (recommended: conda or pip)Main dependencies may include: Python 3.8+ Jupyter Notebook / Lab NumPy, SciPy, Matplotlib PyTorch / TensorFlow (depending on the INR implementation) ERT forward modeling tools (e.g., pyGIMLi, Res2Dinv call interfaces)(Please add a complete list based on actual requirements) Suggested execution order To regenerate forward data: run Forward Modeling ERT (update).ipynb To train the INR model: run Implicit Neural Representation.ipynb For traditional methods: extract Numerical.zip and ensure output file paths are correct Final plotting: run Figures and Analysis of ERT-3D-INR paper.ipynb Important NotesThis repository does not include raw measured data (if applicable). All synthetic data are generated via forward modeling or predefined models. The .ipynb files depend on specific file paths; please keep the default directory structure. To reproduce all computations, run the notebooks in the suggested order and ensure intermediate results are saved properly. CitationIf you use the data or code from this repository, please cite the following paper (to be filled): Authors, Title, Journal/Conference, Year.各文件详细说明 1. INR.zip 文件内容：内含隐式神经表征（Implicit Neural Representation，INR）方法所用的原始数据（如输入坐标、观测数据），以及训练完成后得到的对应模型输出（如重建电阻率场、预测值）。 格式：ZIP压缩包；解压后典型文件格式包括.npy、.mat、.txt、.csv或自定义格式。 用途：用于复现INR模型的预测结果，或作为进一步分析的输入数据。 2. Numerical.zip 文件内容：内含Res2Dinv与Res3Dinv方法的原始数据及模型输出（如合成/野外电阻率层析成像（Electrical Resistivity Tomography，ERT）数据、反演得到的电阻率分布）。 格式：ZIP压缩包；内部文件结构遵循对应方法的标准输入输出格式。 用途：与INR结果进行对比，以验证所提出的方法。 3. Implicit Neural Representation.ipynb 类型：Jupyter Notebook 功能：实现隐式神经表征（INR）方法的完整训练流程，涵盖数据加载、网络定义、损失函数、训练循环以及模型保存环节。 输出：训练日志、损失曲线、中间结果与最终模型权重（通常保存于INR.zip内或单独的文件夹中）。 依赖项：详见下文"环境要求"部分。 4. Forward Modeling ERT (update).ipynb 类型：Jupyter Notebook 功能：执行ERT正演模拟，以根据给定的电阻率模型计算电位/视电阻率响应；包含更新后的参数设置、求解器配置与结果可视化功能。 输出：正演响应数据、网格信息、合成数据集（可作为INR.zip或Numerical.zip中的输入数据使用）。 适用性：可作为INR方法与传统反演方法的通用数据生成来源。 5. Figures and Analysis of ERT-3D-INR paper.ipynb 类型：Jupyter Notebook 功能：生成《ERT-3D-INR》论文正文中的全部图表（如对比图、误差分布图、剖面图、收敛曲线），并对结果进行统计分析。 输入：读取INR.zip、Numerical.zip及正演建模脚本生成的输出文件。 输出：可导出的.png、.pdf或矢量图形文件，与论文中展示的图表完全一致。 使用指南 1. 下载并解压：将INR.zip与Numerical.zip解压至当前工作目录（若脚本支持自动读取，也可保留压缩包状态）。 2. 配置运行环境（推荐使用conda或pip进行环境管理）： 主要依赖项包括： - Python 3.8及以上版本 - Jupyter Notebook / Lab - NumPy、SciPy、Matplotlib - PyTorch / TensorFlow（取决于具体的INR实现方案） - ERT正演建模工具（如pyGIMLi、Res2Dinv调用接口）（请根据实际需求补充完整依赖列表） 3. 推荐执行顺序： - 若需重新生成正演数据：运行Forward Modeling ERT (update).ipynb - 若需训练INR模型：运行Implicit Neural Representation.ipynb - 若需使用传统反演方法：解压Numerical.zip并确保输出文件路径正确 - 最终绘图与分析：运行Figures and Analysis of ERT-3D-INR paper.ipynb 重要提示 1. 本仓库未包含原始实测数据（如适用），所有合成数据均通过正演模拟或预定义模型生成。 2. 所有.ipynb文件均依赖特定的文件路径，请保持默认目录结构。 3. 若需复现全部计算流程，请按照推荐顺序运行脚本，并确保中间结果已正确保存。 引用说明 若您使用本仓库中的数据或代码，请引用以下论文（待补充）： 作者，论文标题，期刊/会议名称，年份。