Feature Extraction

Bottleneck Features

Compressed representations extracted just before the classification layers of a pre-trained network. These bottleneck features capture essential semantic information in a compact format ideal for transfer learning.

← 뒤로