方法:embedding module 提取特征,然后通过Bi-Similarity module双相似度模块来计算得分然后取均值
数据集: FGVC-Aircraft, Stanford-Cars, Stanford-Dogs and CUB-200-2011
类别:train:val:test = 2:1:1
尺寸:84*84,不使用边框
trained from scratch
backbone: Conv4, Conv64F
结果
方法:embedding module 提取特征,然后通过Bi-Similarity module双相似度模块来计算得分然后取均值
数据集: FGVC-Aircraft, Stanford-Cars, Stanford-Dogs and CUB-200-2011
类别:train:val:test = 2:1:1
尺寸:84*84,不使用边框
trained from scratch
backbone: Conv4, Conv64F
结果