Deep learning from multiple crowds: a case study of humanitarian mapping

Satellite images are widely applied in humanitarian mapping that labels buildings, roads, and so on for humanitarian aid and economic development. However, the labeling now is mostly done by volunteers. In this paper, we utilize deep learning to solve humanitarian mapping tasks of a mobile software...

Full description

Saved in:
Bibliographic Details
Main Authors: Chen, Jiaoyan (Author) , Zipf, Alexander (Author)
Format: Article (Journal)
Language:English
Published: 2019
In: IEEE transactions on geoscience and remote sensing
Year: 2018, Volume: 57, Issue: 3, Pages: 1713-1722
ISSN:1558-0644
Online Access: Get full text
Author Notes:Jiaoyan Chen, Yan Zhou, Alexander Zipf, and Hongchao Fan

MARC

LEADER 00000caa a2200000 c 4500
001 1666566497
003 DE-627
005 20220816163755.0
007 cr uuu---uuuuu
008 190531r20192018xx |||||o 00| ||eng c
024 7 |a 10.1109/TGRS.2018.2868748  |2 doi 
035 |a (DE-627)1666566497 
035 |a (DE-599)KXP1666566497 
035 |a (OCoLC)1341225951 
040 |a DE-627  |b ger  |c DE-627  |e rda 
041 |a eng 
084 |a 31  |2 sdnb 
100 1 |a Chen, Jiaoyan  |e VerfasserIn  |0 (DE-588)1187539295  |0 (DE-627)1666566799  |4 aut 
245 1 0 |a Deep learning from multiple crowds  |b a case study of humanitarian mapping  |c Jiaoyan Chen, Yan Zhou, Alexander Zipf, and Hongchao Fan 
264 1 |c 2019 
300 |a 10 
336 |a Text  |b txt  |2 rdacontent 
337 |a Computermedien  |b c  |2 rdamedia 
338 |a Online-Ressource  |b cr  |2 rdacarrier 
500 |a Date of Publication: 03 October 2018 
500 |a Gesehen am 31.05.2019 
520 |a Satellite images are widely applied in humanitarian mapping that labels buildings, roads, and so on for humanitarian aid and economic development. However, the labeling now is mostly done by volunteers. In this paper, we utilize deep learning to solve humanitarian mapping tasks of a mobile software named MapSwipe. The current deep learning techniques, e.g., convolutional neural network (CNN), can recognize ground objects from satellite images but rely on numerous labels for training for each specific task. We solve this problem by fusing multiple freely accessible crowdsourced geographic data and propose an active learning-based CNN training framework named MC-CNN to deal with the quality issues of the labels extracted from these data, including incompleteness (e.g., some kinds of object are not labeled) and heterogeneity (e.g., different spatial granularities). The method is evaluated with building mapping in South Malawi and road mapping in Guinea with level-18 satellite images provided by Bing Map and volunteered geographic information from OpenStreetMap, MapSwipe, and OsmAnd. The results based on multiple metrics, including Precision, Recall, F1 Score, and area under the receiver operating characteristic curve, show that MC-CNN can fuse the crowdsourced labels for higher prediction performance and be successfully applied in MapSwipe for humanitarian mapping with 85% labor saved and an overall accuracy of 0.86 achieved. 
534 |c 2018 
650 4 |a Active learning 
650 4 |a active learning-based CNN training framework 
650 4 |a Bing Map 
650 4 |a building mapping 
650 4 |a Buildings 
650 4 |a convolutional neural nets 
650 4 |a convolutional neural network 
650 4 |a crowdsourced labels 
650 4 |a crowdsourcing 
650 4 |a deep learning 
650 4 |a economic development 
650 4 |a geographic information systems 
650 4 |a ground objects 
650 4 |a humanitarian aid 
650 4 |a humanitarian mapping 
650 4 |a humanitarian mapping tasks 
650 4 |a Labeling 
650 4 |a learning (artificial intelligence) 
650 4 |a level-18 satellite images 
650 4 |a Machine learning 
650 4 |a MapSwipe 
650 4 |a MC-CNN 
650 4 |a mobile computing 
650 4 |a mobile software 
650 4 |a multiple crowds 
650 4 |a multiple freely accessible crowdsourced geographic data 
650 4 |a multiple metrics 
650 4 |a numerous labels 
650 4 |a OpenStreetMap 
650 4 |a OsmAnd 
650 4 |a road mapping 
650 4 |a Roads 
650 4 |a satellite image 
650 4 |a Satellites 
650 4 |a Task analysis 
650 4 |a Training 
650 4 |a volunteered geographic information (VGI) 
700 1 |a Zipf, Alexander  |d 1971-  |e VerfasserIn  |0 (DE-588)123246369  |0 (DE-627)082437076  |0 (DE-576)175641056  |4 aut 
773 0 8 |i Enthalten in  |a Institute of Electrical and Electronics Engineers  |t IEEE transactions on geoscience and remote sensing  |d New York, NY : IEEE, 1964  |g 57(2019), 3, Seite 1713-1722  |h Online-Ressource  |w (DE-627)324487967  |w (DE-600)2027520-1  |w (DE-576)094085927  |x 1558-0644  |7 nnas 
773 1 8 |g volume:57  |g year:2019  |g number:3  |g pages:1713-1722  |g extent:10  |a Deep learning from multiple crowds a case study of humanitarian mapping 
951 |a AR 
992 |a 20190531 
993 |a Article 
994 |a 2019 
998 |g 123246369  |a Zipf, Alexander  |m 123246369:Zipf, Alexander  |d 120000  |d 120700  |e 120000PZ123246369  |e 120700PZ123246369  |k 0/120000/  |k 1/120000/120700/  |p 3 
999 |a KXP-PPN1666566497  |e 3480380220 
BIB |a Y 
SER |a journal 
JSO |a {"name":{"displayForm":["Jiaoyan Chen, Yan Zhou, Alexander Zipf, and Hongchao Fan"]},"language":["eng"],"type":{"bibl":"article-journal","media":"Online-Ressource"},"recId":"1666566497","title":[{"title_sort":"Deep learning from multiple crowds","subtitle":"a case study of humanitarian mapping","title":"Deep learning from multiple crowds"}],"origin":[{"dateIssuedDisp":"2019","dateIssuedKey":"2019"}],"relHost":[{"origin":[{"dateIssuedDisp":"1964-","publisher":"IEEE","dateIssuedKey":"1964","publisherPlace":"New York, NY"}],"id":{"eki":["324487967"],"issn":["1558-0644"],"zdb":["2027520-1"]},"titleAlt":[{"title":"Transactions on geoscience and remote sensing"}],"physDesc":[{"extent":"Online-Ressource"}],"note":["Gesehen am 31.07.25"],"part":{"year":"2019","text":"57(2019), 3, Seite 1713-1722","pages":"1713-1722","extent":"10","issue":"3","volume":"57"},"language":["eng"],"type":{"bibl":"periodical","media":"Online-Ressource"},"corporate":[{"roleDisplay":"VerfasserIn","role":"aut","display":"Institute of Electrical and Electronics Engineers"}],"pubHistory":["Volume 18, number 1 (January 1980)-"],"recId":"324487967","title":[{"title":"IEEE transactions on geoscience and remote sensing","title_sort":"IEEE transactions on geoscience and remote sensing","subtitle":"a publication of the IEEE Geoscience and Remote Sensing Society"}],"disp":"Institute of Electrical and Electronics EngineersIEEE transactions on geoscience and remote sensing"}],"id":{"eki":["1666566497"],"doi":["10.1109/TGRS.2018.2868748"]},"physDesc":[{"extent":"10 S."}],"note":["Date of Publication: 03 October 2018","Gesehen am 31.05.2019"],"person":[{"given":"Jiaoyan","display":"Chen, Jiaoyan","role":"aut","family":"Chen","roleDisplay":"VerfasserIn"},{"family":"Zipf","roleDisplay":"VerfasserIn","given":"Alexander","role":"aut","display":"Zipf, Alexander"}]} 
SRT |a CHENJIAOYADEEPLEARNI2019