Meta tags:
description=Novel training methods that exploit the spatio-temporal structure of remote sensing data.;
keywords=satellite imagery, remote sensing, self-supervised learning, representation learning, geo-location, object detection, classification, segmentation;

Headings (most frequently used words):

the, in, of, shows, images, we, geoimagenet, learning, over, time, fmow, dataset, each, image, changes, for, our, distribution, geography, aware, self, supervised, concept, functional, map, world, metadata, associated, with, is, shown, underneath, can, see, contrast, brightness, cloud, cover, etc, these, render, spatially, aligned, useful, constructing, additional, positives, abstract, model, top, original, moco, v2, framework, bottom, schematic, overview, approach, some, examples, from, below, list, their, latitudes, longitudes, city, country, name, study, use, latitude, and, longitude, information, unsupervised, poster, bibtex,

Text of the page (most frequently used words):
the (41), and (21), #learning (15), that (10), images (10), supervised (9), geo (9), our (8), image (8), classification (7), this (6), website (6), geography (6), aware (6), for (6), dataset (6), show (6), can (6), from (5), tasks (5), contrastive (5), gap (5), self (4), kumar (4), shows (4), geoimagenet (4), tagged (4), imagenet (4), with (4), their (4), proposed (4), improve (4), over (4), time (4), temporal (4), between (4), object (4), detection (4), semantic (4), segmentation (4), remote (4), sensing (4), data (4), you (3), are (3), code (3), fmow (3), approach (3), able (3), some (3), visual (3), cues (3), picture (3), was (3), indian (3), where (3), world (3), moco (3), spatially (3), aligned (3), accuracy (3), method (3), which (2), ayush (2), uzkent (2), burak (2), meng (2), chenlin (2), tanmay (2), burke (2), marshall (2), lobell (2), david (2), ermon (2), stefano (2), iccv (2), 2021 (2), distribution (2), demonstrate (2), associated (2), coordinates (2), more (2), located (2), performance (2), application (2), examples (2), have (2), see (2), captured (2), elephant (2), african (2), model (2), other (2), each (2), unsupervised (2), construct (2), positive (2), pairs (2), location (2), design (2), pre (2), text (2), close (2), datasets (2), experiments (2), significantly (2), top (2), cover (2), functional (2), map (2), methods (2), paper (2), changes (2), means, free, borrow, just, ask, link, back, page, footer, please, remember, remove, analytics, included, header, not, want, your, source, licensed, under, creative, commons, attribution, sharealike, international, license, format, borrowed, project, nerfies, article, ayush2021geography, title, author, journal, year, bibtex, poster, further, effectiveness, searched, using, flickr, api, were, find, 543, 435, lat, lon, across, 5150, class, categories, challenging, than, highly, imbalanced, contains, about, classes, extend, approaches, showing, potential, any, figure, above, predicted, example, person, sombrerohat, mexico, similarly, india, there, large, population, elephants, next, larger, size, trained, predict, taken, should, identify, transferable, differentiate, ele, phants, counterparts, below, list, latitudes, longitudes, city, country, name, study, use, latitude, longitude, information, leveraging, consisting, high, spatial, resolution, satellite, baseline, particular, when, testing, learned, representations, miou, land, interestingly, even, outperform, counterpart, target, applications, utilizing, original, framework, schematic, overview, bottom, narrowed, computer, vision, explore, unlabeled, often, abundant, but, labeled, scarce, first, due, different, characteristics, non, trivial, persists, standard, benchmarks, propose, novel, training, exploit, spatio, structure, leverage, closes, moreover, also, beapplied, improving, down, stream, various, abstract, concept, metadata, shown, underneath, contrast, brightness, cloud, etc, these, render, useful, constructing, additional, positives, arxiv, stanford, university,

Text of the page (random words):
geography aware self supervised learning geography aware self supervised learning kumar ayush burak uzkent chenlin meng kumar tanmay marshall burke david b lobell stefano ermon stanford university iccv 2021 paper arxiv code dataset images over time concept in the functional map of the world fmow dataset the metadata associated with each image is shown underneath we can see changes in contrast brightness cloud cover etc in the images these changes render spatially aligned images over time useful for constructing additional positives abstract contrastive learning methods have significantly narrowed the gap between supervised and unsupervised learning on computer vision tasks in this paper we explore their application to geo located datasets e g remote sensing where unlabeled data is often abundant but labeled data is scarce we first show that due to their different characteristics a non trivial gap persists between contrastive and supervised learning on standard benchmarks to close the gap we propose novel training methods that exploit the spatio temporal structure of remote sensing data we leverage spatially aligned images over time to construct temporal positive pairs in contrastive learning and geo location to design pre text tasks our experiments show that our proposed method closes the gap between contrastive and supervised learning on image classification object detection and semantic segmentation for remote sensing moreover we demonstrate that the proposed method can also beapplied to geo tagged imagenet images improving down stream performance on various tasks model top shows the original moco v2 framework bottom shows the schematic overview of our approach by leveraging spatially aligned images over time to construct temporal positive pairs in contrastive learning and geo location in the design of pre text tasks we are able to close the gap between self supervised and supervised learning on image classification object detection and semantic segmentation on remote sensing and other geo tagged image datasets our experiments on the functional map of the world fmow dataset consisting of high spatial resolution satellite images show that we improve moco v2 baseline significantly in particular we improve it by 8 classification accuracy when testing the learned representations on image classification 2 ap on object detection 1 miou on semantic segmentation and 3 top 1 accuracy on land cover classification interestingly our geography aware learning can even outperform the supervised learning counterpart on temporal data classification by 2 with the proposed method we can improve the accuracy on target applications utilizing object detection and semantic segmentation geoimagenet some examples from geoimagenet dataset below each image we list their latitudes longitudes city country name in our study we use the latitude and longitude information for unsupervised learning to further demonstrate the effectiveness of our geography aware learning approach we searched for geo tagged images in imagenet using the flickr api and were able to find 543 435 images with their associated coordinates lat i lon i across 5150 class categories this dataset is more challenging than imagenet 1k as it is highly imbalanced and contains about 5x more classes we extend the proposed approaches to geo located imagenet and show that geography aware learning can improve the performance of moco v2 by 2 on image classification showing the potential application of our approach to any geo tagged dataset we show some examples from geoimagenet in the figure above for some images we have geo coordinates that can be predicted from visual cues for example we see that a picture of a person with a sombrerohat was captured in mexico similarly an indian elephant picture was captured in india where there is a large population of indian elephants next to it we show the picture of an african elephant which is larger in size if a model is trained to predict where in the world the image was taken it should be able to identify visual cues that are transferable to other tasks e g visual cues to differentiate indian ele phants from the african counterparts shows the distribution of the fmow shows the distribution of geoimagenet poster bibtex article ayush2021geography title geography aware self supervised learning author ayush kumar and uzkent burak and meng chenlin and tanmay kumar and burke marshall and lobell david and ermon stefano journal iccv year 2021 the format of the website is borrowed from the nerfies project website this website is licensed under a creative commons attribution sharealike 4 0 international license this means you are free to borrow the source code of this website we just ask that you link back to this page in the footer please remember to remove the analytics code included in the header of the website which you do not want on your website