CNN Image Retrieval Learns from BoW: Unsupervised Fine-Tuning with Hard Examples

8 Apr 2016Filip RadenovićGiorgos ToliasOndřej Chum

Convolutional Neural Networks (CNNs) achieve state-of-the-art performance in many computer vision tasks. However, this achievement is preceded by extreme manual annotation in order to perform either training from scratch or fine-tuning for the target task... (read more)

PDF Abstract

Evaluation results from the paper

Task Dataset Model Metric name Metric value Global rank Compare
Image Retrieval Oxf105k siaMAC+QE* MAP 77.9% # 5
Image Retrieval Oxf5k siaMAC+QE* MAP 82.9% # 5
Image Retrieval Par106k siaMAC+QE* mAP 78.3% # 5
Image Retrieval Par6k siaMAC+QE* mAP 85.6% # 4