Whenever you're ready, walk me through how you'd approach this.
How do you approach feature extraction in image processing?
Explain how you would preprocess images, choose between handcrafted descriptors and learned representations, prevent information leakage, and evaluate whether the extracted features support the downstream task. Include practical considerations such as scale, rotation, lighting variation, noise, computational cost, and reproducibility.