Pedestrian_Alignment
Pedestrian_Alignment copied to clipboard
Relationship to another article
Hi,i want to know if this network architecture is improved on the previous network architecture in the artical "A Discriminatively Learned CNN Embedding for Person Re-identification"?I noticed a lot of similarities.
Yes. Because the best model (in the paper) is with the spirit of "A Discriminatively Learned CNN Embedding for Person Re-identification" and GAN. The verification loss and identification loss are added.