踏实可乐

文章
5
资源
0
加入时间
2年10月21天

Cross-Domain Few-Shot Classification VIA Learned Feature-Wise Transformation 论文解读摘要简介:方法:FEATURE-WISE TRANSFORMATION LAYER:LEARNING THE FEATURE-WISE TRANSFORMATION LAYERS :实验:

Cross-Domain Few-Shot Classification VIA Learned Feature-Wise Transformation摘要简介:方法:FEATURE-WISE TRANSFORMATION LAYER:LEARNING THE FEATURE-WISE TRANSFORMATION LAYERS :实验:摘要简介:小样本分类的目的是在每一个类别仅有少量标注样本的情况下识别新的类别。当前metric-based(基于度量)的小样本分类方法是:先利用metric f