Research On Classification System Construction And Retrieval Of Drawing Image Learning Resources

The publication and implementation of the “Action Plan for Educational Informatization 2.0 ”marks that educational informatization has entered a new stage of development.It is imperative to expand and improve the public service system of digital education resources,promote open resource pooling and sharing,collect and converge abundant teaching,scientific research and cultural resources on the Internet by using big data technology,and provide massive and appropriate learning resources services for schools at all levels and all learners.Throughout the existing digital learning resources and resource service systems,the types and quantities of resources have been relatively rich,but the quality and practical application are not satisfactory.Especially text,sound,graphics,images,animation,video and other multimedia materials are scattered,uneven and disordered,which makes it inconvenient for learners and teachers to obtain and utilize.As an important information carrier,graphics and images have the characteristics of intuitive expression,rich content,no language restriction and easy to understand and disseminate.They are the important content of multimedia information and an important part of digital learning resources.Especially,the image resources drawn by paper,pencil,computer and other assistant tools make the objective things or abstract concepts which are inconvenient or impossible to observe directly in the real world intuitively presented.It has important application value in the process of education and teaching.In this paper,the drawing image resources are taken as the research object,and the construction of its classification system and convenient and efficient retrieval are studied,so as to integrate the high-quality digital image resources in the Internet environment,and to provide a green,shared,data-rich and query-convenient drawing image learning resource platform for educators and learners.The main research work is as follows:(1)Theoretical research.On the basis of consulting,sorting out and analyzing relevant literature data,this paper makes a thorough analysis of the types and characteristics of drawing image learning resources and their application value in the process of education,the organization and classification methods of network information resources,image classification technology based on convolutional neural network and information retrieval technology based on Lucene,so that the strategy and method of resource classification and platform development are clarified.(2)Practical research.On the basis of the existing research,the classification system of drawing image learning resources is constructed,and the classification system is coded by parallel codes with digital format,fixed length and increasing order.The CNN model is trained by sample set of image resources those acquired by spider technology,so as to realize automatic classification and labeling of large-scale image resources,and then construct a drawing image learning resource database and develop a resource retrieval platform.(3)Investigation research.Based on Technology Acceptance Model,combined with the cognitive and competence level of different user groups,a questionnaire for middle school students,college students and teachers was designed to investigate the application status of mapping image resources and the practical application effect of the resource retrieval platform constructed in this study.The problems in the process of resource application and the advantages of the resource retrieval platform were summarized.It points out the direction for the follow-up research work.The educational significance of drawing image learning resources has been fully affirmed by students and teachers,and the resource retrieval platform designed and developed in this study has also been widely recognized by users.However,due to the limited time and technical ability of researchers,there are still many problems and shortcomings in the resource retrieval platform.Future research work will focus on expanding training sample set to modify CNN classification model,improving image classification system,and achieving convenient and accurate retrieval of large-scale image database.