English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

A deep convolutional neural network (CNN) has been widely used in imageclassification and gives better classification accuracy than the othertechniques. The softmax cross-entropy loss function is often used forclassification tasks. There are some works to introduce the additional terms inthe objective function for training to make the features of the output layermore discriminative. The neuron-wise discriminant criterion makes the inputfeature of each neuron in the output layer discriminative by introducing thediscriminant criterion to each of the features. Similarly, the center loss wasintroduced to the features before the softmax activation function for facerecognition to make the deep features discriminative. The ReLU function isoften used for the network as an active function in the hidden layers of theCNN. However, it is observed that the deep features trained by using the ReLUfunction are not discriminative enough and show elongated shapes. In thispaper, we propose to use the neuron-wise discriminant criterion at the outputlayer and the center-loss at the hidden layer. Also, we introduce the onlinecomputation of the means of each class with the exponential forgetting. Wenamed them adaptive neuron-wise discriminant criterion and adaptive centerloss, respectively. The effectiveness of the integration of the adaptiveneuron-wise discriminant criterion and the adaptive center loss is shown by theexperiments with MNSIT, FashionMNIST, CIFAR10, CIFAR100, and STL10. Source codeis at https://github.com/i13abe/Adaptive-discriminant-and-center

Adaptive Neuron-wise Discriminant Criterion and Adaptive Center Loss at Hidden Layer for Deep Convolutional Neural Network