基于张量特征的小样本图像快速分类方法

张艳莎, 冯夫健, 王杰, 潘凤, 谭棉, 张再军, 王林

Tensor feature⁃based faster classification network for few⁃shot learning

Yansha Zhang, Fujian Feng, Jie Wang, Feng Pan, Mian Tan, Zaijun Zhang, Lin Wang

表1 小样本图像分类算法的对比

Table 1 Comparison of few?shot image classification algorithms

Method Backbone miniImageNet CUB CIFAR⁃FS 1⁃shot 5⁃shot 1⁃shot 5⁃shot 1⁃shot 5⁃shot IDeMet⁃Net[9] ResNet⁃18 59.14±0.86 74.63±0.74 - - - - AFHN[12] ResNet⁃18 62.38±0.72 78.16±0.56 70.53±1.01 83.95±0.63 68.32±0.93 81.45±0.87 VI⁃Net[14] ResNet⁃18 61.05 78.60 74.76 86.84 - - CFA⁃PN[18] ResNet⁃12 60.47±0.61 77.82±0.44 Dual TriNet[19] ResNet⁃18 58.80±1.37 76.71±0.69 69.61 84.10 63.41±0.64 78.43±0.64 EDANet+ PrototypicalNet[20] ResNet⁃50 63.35 79.74 - - - - TFFCN ResNet⁃18 63.99±0.80 79.95±0.61 75.13±0.86 88.08±0.47 73.00±0.95 86.52±0.62