With static photography as the main reflection method,photo archives are memory records with preservation value directly formed by national institutions,social organizations,and individuals in social activities. Due to the quantity,quickly retrieving related photos from a large number of photos according to some special conditions (such as location,key festives,character combinations,etc. ) is difficult . In addition,using online photo management software is inconvenient due to the particularity of archive photos. Thus,a local photo archive management system utilizing the multimodal retrieval tech-nology to quickly retrieve the qualified photos from a large number of photos according to the conditions such as location,key festivals,and character combinations is introduced in this paper. With this sys-tem,process of photo uploading and screening will be simplified to quickly find out the photos satisfying specific conditions,greatly easing the workload of photo archive administrators.
关键词
多模态检索/光学字符识别/人脸识别
Key words
multimodal retrieval/optical character recognition(OCR)/facial recognition
维普
万方数据