Preprocessing refers to a series of deterministic steps applied to all images (including those for training, validation, testing, and production) before they are input into the model. These steps aim to standardize image formats, enhance visual features, or reduce noise, ensuring the model can process the data more effectively.
