Panorama to panorama matching for location recognition
Authors:
- Ahmet IscenAIAhmet Iscen
- Giorgos ToliasGTGiorgos Tolias
- Yannis AvrithisYAYannis Avrithis
- Teddy FuronTFTeddy Furon
- Ondrej ChumOCOndrej Chum
Preprint
Abstract
Location recognition is commonly treated as visual instance retrieval on "street view" imagery. The dataset items and queries are panoramic views, i.e. groups of images taken at a single location. This work introduces a novel panorama-to-panorama matching process, either by aggregating features of individual images in a group or by explicitly constructing a larger panorama. In either case, multiple views are used as queries. We reach near perfect location recognition on a standard benchmark with only four query views.Keywords
All Related Versions
This publication has 25 references indexed in Scilit:
- Cross-Dimensional Weighting for Aggregated Deep Convolutional Features2016
- CNN Image Retrieval Learns from BoW: Unsupervised Fine-Tuning with Hard Examples2016
- Deep Image Retrieval: Learning Global Representations for Image Search2016
- DisLocation: Scalable Descriptor Distinctiveness for Location Recognition2015
- Graph-Based Discriminative Learning for Location RecognitionInternational Journal of Computer Vision, 2014
- Neural Codes for Image Retrieval2014
- Improving Image-Based Localization by Active Correspondence Search2012
- City-Scale Location Recognition2007
- Automatic Panoramic Image Stitching using Invariant FeaturesInternational Journal of Computer Vision, 2006
- Image Based Localization in Urban Environments2006