全域性特徵

Without Fine-tuning

• Neural codes for image retrieval, ECCV, 2014
• Multi-scale orderlesspooling of deep convolutional activation features, ECCV, 2014 (VLAD coding)
• CNN Features off-the-shelf: an Astounding Baseline for Recognition, CVPR, 2014
• Aggregating Deep Convolutional Features for Image Retrieval, ICCV, 2015
• Exploiting Local Features from Deep Networks for Image Retrieval, CVPR, 2015, (VLAD coding)
• Bags of Local Convolutional Features for Scalable Instance Search, ICMR, 2016
• MAC: Particular object retrieval with integral max-pooling of CNN activations, ICLR, 2016
• Cross-dimensional Weighting for Aggregated Deep Convolutional Features, ECCV, 2016

With Fine-tuning

• Fisher Vectors Meet Neural Networks: A Hybrid Classification Architecture, CVPR, 2015 (FV coding)
• NetVLAD: CNN architecture for weakly supervised place recognition, CVPR, 2016
• Faster R-CNN Features for Instance Search, CVPR, 2016
• R-MAC: Deep Image Retrieval: Learning global representations for image search, ECCV, 2016
• R-MAC: End-to-end Learning of Deep Visual Representations for Image Retrieval, IJCV, 2017 (detailed description of R-MAC)

區域性特徵

• Local convolutional features with unsupervised training for image retrieval, ICCV, 2015
• Largescale image retrieval with attentive deep local features, ICCV, 2017

新評價方案

• Revisiting Oxford and Paris: Large-Scale Image Retrieval Benchmarking, CVPR, 2018