A Discriminative Siamese Network with Softmax loss.

💪 💪 💪 💪 💪

Generate Dataset

Generate two-patchs dataset randomly and classification dataset, including CASIA-WebFace trained dataset and lfw test dataset. Detailly, all network are trained and tested on the same dataset, but for Siamese Network it is two-patchs, and for Classification Network one-patch.

Training on CASIA-WebFace, and test on lfw.

Trace

1. Softmax_loss

   • CaffeNet Off-the-shelf
   • AlexNet Off-the-shelf
   • CaffeNet Fine-tuning
   • AlexNet Fine-tuning

2. Softmax_loss + Center_loss

   • CaffeNet Off-the-shelf
   • AlexNet Off-the-shelf
   • CaffeNet Fine-tuning
   • AlexNet Fine-tuning

3. Contrastive_loss on Siamese Network

   • CaffeNet Off-the-shelf
   • AlexNet Off-the-shelf

4. 2-Channels Network

   • CaffeNet Off-the-shelf
   • AlexNet Off-the-shelf

5. A Discriminative Siamese Network with Softmax loss

   • CaffeNet Off-the-shelf
   • AlexNet Off-the-shelf