Abstract：MetaEnhanceMetadata Quality Improvement for Electronic Theses and Dissertations of University Libraries Metadata quality is crucial for discovering digital objects through digital libraryDLinterfacesHoweverdue to various reasonsthe metadata of digital objects often exhibits incompleteinconsistentand incorrect valuesWe investigate methods to automatically detectcorrectand canonicalize scholarly metadatausing seven key fields of electronic theses and dissertationsETDsas a case studyWe propose MetaEnhancea framework that utilizes state-of-the-art artificial intelligenceAImethods to improve the quality of these fieldsTo evaluate MetaEnhancewe compiled a metadata quality evaluation benchmark containing 500 ETDsby combining subsets sampled using multiple criteriaWe evaluated MetaEnhance against this benchmark and found that the proposed methods achieved nearly perfect F1-scores in detecting errors and F1-scores ranging from 085 to 100 for correcting five of seven key metadata fieldsThe codes and data are publicly available on GitHub 1 1 and IndexingThrough the improvement of metadatathe detection of digital library objects can be improvedInstificational CollectionETDS is usually custody by college libraries or centralized online reservoirssuch as PROQUESTso it involves the management of institutional collections and improves its metadata qualityIn this studythey use the method of natural language processingNLPand computer visionCVto improve the quality of meta dataand by automatic detectioncorrectionand standardized metadataThis is metadata that uses NLP technology to process and improve digital objects