No description, website, or topics provided.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
README.md
grids.txt grids May 11, 2017
predict_geolocation.ipynb use updated model May 19, 2017
predict_places365.ipynb

README.md

Geo-location Tutorial

Download model: https://s3.amazonaws.com/mmcommons-tutorial/models/RN101-5k500-0012.params https://s3.amazonaws.com/mmcommons-tutorial/models/RN101-5k500-symbol.json

Geolocation model inspired by ideas presented in: PlaNet - Photo Geolocation with Convolutional Neural Networks (ECCV 2016), Tobias Weyand, Ilya Kostrikov, James Philbin https://research.google.com/pubs/pub45488.html

Data and Classes

Our data come from the geotagged images in the YFCC100M Multimedia Commons dataset. Training, validation, and test images are split so that images uploaded by the same person do not appear in multiple sets. Classes are created with the training data using Google's S2 Geometry Library as described in the PlaNet paper above. The classes are defined in grids.txt where the i-th line is the i-th class and the columns are: S2 Cell Token, Latitude, Longitude.

Difference between our model and PlaNet:

              Ours PlaNet
Dataset source Multimedia Commons Images crawled from the web
Training set 33.9 million 91 million
Validation 1.8 million 34 million
S2 Cell Partitioning t_1=5000, t_2=500 ==> 15,527 cells t_1=10,000, t_2=50 ==> 26,263 cells
Model ResNet-101 GoogleNet
Optimization SGD with Momentum and LR Schedule Adagrad
Training time 9 days on 16 NVIDIA K80 GPUs (p2.16xlarge), 12 epochs 2.5 months on 200 CPU cores
Framework MXNet DistBelief
Test set Placing Task 2016 Test Set (1.5 million Flickr images) 2.3 M geo-tagged Flickr images

Result

Im2GPS test set

The values indicate the percentages of images within test set that were correctly localized within the given distance.

Method 1km 25km 200km 750km 2500km
PlaNet 8.4% 24.5% 37.6% 53.6% 71.3%
Ours 16.8% 39.2% 48.9% 67.9% 82.2%

Flickr Images

Note that these result in the table are not directly comparable as the test set images used in PlaNet is not publicly released. The values indicate the percentages of images within test set that were correctly localized within the given distance.

Method 1km 25km 200km 750km 2500km
PlaNet 3.6% 10.1% 16.0% 28.4% 48.0%
Ours 6.2% 13.5% 20.8% 35.6% 55.2%