rodrigozamb/Data-Wrangling-Mario-Dataset

Data-Wrangling-Mario-Dataset

Very simple project for data wrangling of Mario Bros Datasets.

Dependencies

The following dependencies must be installed to run this project:

• Python3: sudo apt install python3-pip
• Numpy: pip3 install numpy
• Pandas: pip3 install pandas
• Tensorflow: pip3 install tensorflow
• Keras: pip3 install keras
• Cv2: pip3 install opencv-python (needs to upgrade pip: pip3 install --upgrade pip)

Build and Run

Create Config File

The config.json configuration file (under conf/ directory) has the following structure:

{
    "datasetInfo": {
        "inputDataset": "/path/to/source/dataset/files/",
        "outputDataset": "/path/to/output/dataset/files/",
        "classes": ["Class1", "Class2", ..., "ClassN"],
        "chunkSize": chunk_size, (Recommended Value: 6)
        "imageSize": image_size (Recommended Value: 224)
    },
    "featureExtractor": {
        "name": "model_name" (MobileNetV2, ResNet50V2 or VGG16)
    }
}

• datasetInfo/inputDataset: directory to load video data.
• datasetInfo/outputDataset: directory to save image data.
• datasetInfo/classes: classes to extract frames.
• datasetInfo/chunkSize: number of frames that compose the chunk.
• datasetInfo/imageSize: size of frames to use in the execution.
• featureExtractor/name: name of the feature extractor model to use in the execution.

Execution

Steps to execute this code:

• Gather data.
• Create config.json.
• Execute the following command under src/ directory: python3 Main.py