Skip to main content
. 2025 Jan 6;20(1):e0314086. doi: 10.1371/journal.pone.0314086

Fig 1. Model structure.

Fig 1

The encoder part consists of a layer of PLAB to expand the dimension, four layers of DAL with MLP to mine features, MA Pooling to connect Max Pool and Average Pool, and LBRD consisting of Linear, BatchNorm, ReLU, and Dropout.