GO Loss: A Gaussian Distribution-Based Orthogonal Decomposition Loss for Classification

Table 3

Classification accuracy with SampleNet on Fashion-MNIST dataset.

Methods	Remark	Acc. (%)
Soft-max	—	91.56
Center loss		93.26
Ring loss		93.08
LGM loss		92.33
GCPL loss		92.65
GO loss		93.400.02