01483nas a2200301   4500000000100000008004100001653002400042653001700066653001800083653002000101653001500121653002000136653003400156653002400190653001900214653001600233653001100249653001200260653001300272653001900285100001000304700001100314245004100325856015600366300001400522520062500536020002001161                                       d10aContent recognition10aDeep Learing10aDeep learning10aImage retrieval10aImage data10aImage retrieval10aIntangible cultural heritages10aLearning technology10aMatching rules10aOn-machines10aSketch10aThangka10aTibetans10aUnited Nations1 aY. Li1 aX. Liu00aSketch Based Thangka Image Retrieval  uhttps://www.scopus.com/inward/record.uri?eid=2-s2.0-85104664281&doi=10.1109%2fIAEAC50856.2021.9390657&partnerID=40&md5=fcf790f823382a53b5c529327b56cfb1  a2066-20703 aThangka art is a cultural representative of Tibet of China and one of the Intangible Cultural Heritage recognized by the United Nations. It s a unique art form of painting in Tibetan culture. Thangka images contain rich historical, cultural and religious information. There is little research about Thangka in the field of image retrieval and content recognition. In this paper, we propose a Thangka-sketch image data set, and study the matching rule between Thangka image and sketch by using deep learning technology based on machine vision. In the last of this paper, we show the result of our Thangka image retrieval.  a26896621 (ISSN)