I have seen that a center loss is beneficial in computer vision, especially in face recognition. I have tried to understand this concept from the following material
However, I could not understand the concept clearly. If someone can explain with the help of an example, that would be appreciated.
Have a look at https://arxiv.org/pdf/1804.06655.pdf
https://nirovision.com/request-a-demo/?gclid=Cj0KCQiAhZT9BRDmARIsAN2E-J0c2k0DtQVmFvpyzvceRP79F6eOIu7E_ZjKQGDD6n-dXRAkjY66gcIaAntREALw_wcB
https://arxiv.org/pdf/1707.07391.pdf
Deep convolutional neural networks (CNNs) trained with the softmax loss have achieved remarkable successes in a number of close-set recognition problems, e.g. object recognition, action recognition, etc. Unlike these close-set tasks, face recognition is an open-set problem where the testing classes (persons) are usually different from those in training. Specifically, the center loss simultaneously learns a center for each class, and penalizes the distances between the deep features of the images and their corresponding class centers. Training with the center loss enables CNNs to extract the deep features with two desirable properties: inter-class separability and intra-class compactness. In addition, we extend the center loss in two aspects. First, we adopt parameter sharing between the softmax loss and the center loss, to reduce the extra parameters introduced by centers. Second, we generalize the concept of center from a single point to a region in embedding space, which further allows us to account for intra-class variations. The advanced center loss significantly enhances the discriminative power of deep features.
Article A Comprehensive Study on Center Loss for Deep Face Recognition
for each class in the batch, the center loss compute the center of that particular class and then minimize the distance of each sample of the class from this center. Intuitively, we can have compact representation of a class, resulting in a discriminative embedding . You can also look our similar work: Preprint Git Loss for Deep Face Recognition
- Badges
- Science topic
- Similar topics
- Psychology
- Mental Processes
- Learning