Table 1 The variate and meaning of loss function.
From: A vehicle re-identification framework based on the improved multi-branch feature fusion network
Variate | Meaning |
|---|---|
\({N}_{i}\) | The number of vehicle images per batch |
\({N}_{id}\) | The number of vehicle identities |
\({x}_{j}\) | The output of fully connected layer for \(j\) th identity |
\(y\) | The ground truth identity of input vehicle image |
\({A}_{i}\) | Anchor |
\({P}_{i}\) | Positive |
\({N}_{i}\) | Negative |
\(\delta\) | Minimal margin |
\(\lambda\) | Weight |
