Improved 2D preprocessing script

[hamzaMahdi]

Thank you for this repo. I ran into issues with the 2D script so i modified it:
Optimize image and mask processing for improved performance and storage efficiency

1. Multiprocessing: Implemented parallel processing using Python's multiprocessing module, significantly reducing execution time by utilizing all available CPU cores.

2. Image data type optimization: Changed from implicit float32 (4 bytes/pixel) to explicit uint8 (1 byte/pixel), reducing image data storage by approximately 75%.

3. Compressed NPZ format: Implemented np.savez_compressed() instead of np.save(), further reducing file sizes through compression.

4. Single file storage: Combined image and mask data into a single compressed NPZ file, reducing overhead and improving file management.

These changes significantly reduce storage requirements, especially for large datasets. For example, a 1024x1024 RGB image that previously required about 12 MB (4 bytes * 1024 * 1024 * 3) now requires only about 3 MB (1 byte * 1024 * 1024 * 3), before additional compression from the NPZ format.