Feature Extraction

Deep Convolutional Layers

Initial layers of a convolutional neural network that capture low-level features like edges, textures, and fundamental geometric shapes. These layers are typically reused as-is during transfer learning due to their ability to extract universal patterns.

← Indietro