Development And Application Of Training Practice System Based On Image Recognition Algorithm

Practical teaching is an important part of the whole teaching activities,the proportion of practical teaching hours in the total teaching hours becomes higher and higher,which is an important teaching part.In practice,it was found that the training practice teaching scene could not be well combined with the existing information-based teaching platforms at home and abroad,the promotion process of information-based teaching was slow,and the Internet assisted teaching means could not be used to improve the teaching efficiency.Therefore,it is of practical significance to study the construction of information platform and auxiliary teaching in the context of practical training teaching.Facing the actual needs of the training and practical teaching scene,this paper studies the characteristics and difficulties of the scene from the perspectives of engineering design and algorithm research,and mainly completes the following work:(1)After analyzing the actual needs of the training and practical teaching scene,develop the online teaching system and mobile app suitable for the scene.In view of the difficulty of using the platform by teachers,develop app and specific route jump rules to improve practicability and convenience;Aiming at the problem of difficult homework correction,develop the auxiliary scoring module to provide the ability of homework auxiliary correction;Aiming at the problem of difficult management caused by personnel mobility,adopt the idea of group teaching and develops the team module.(2)Based on the convolution neural network,combined with the complementarity of multi angle image information of the same object and the importance of receptive field in extracting more features of the image,an algorithm model MVAMDCNN combining multi view and multi receptive field features is proposed.(3)After experimental verification,on the public Modelnet40 dataset,the MVAMDCNN algorithm model in this paper is compared with VGG11 in Acc,Precision,Recall and F1＿Score has increased by more than 5% in several evaluation indicators,up to 8%;In the artificially constructed training and practical teaching scene dataset,the evaluation indicators can be improved up to 1.8%;On the artificially constructed job score interval prediction scene data set,the evaluation indicators improvement is up to 0.8%,which proves the effectiveness and superiority of the MVAMDCNN algorithm model in this paper.(4)This paper designs and implements an independent auxiliary scoring service,which is equipped with the trained MVAMDCNN algorithm model proposed in this paper,and provides auxiliary teaching ability through integration with the system.(5)This paper tests the above work.Firstly,the online teaching system and mobile app are tested respectively to demonstrate the functional integrity and practicability of each part.In addition,the integration between the system and services is tested by simulating the actual application scenario from the perspective of function and performance.The test results show that the two part can carry out the auxiliary teaching in the practical teaching scenario through the integration,and 95% of the teachers’ homework correction time can be reduced through the auxiliary correction of homework.