交互式深度学习计算机视觉理论教学研究

邓 丽珍, 张 冉冉
南京邮电大学通信与信息工程学院

摘要

为了让学生更好的理解并掌握基于深度学习的计算机视觉，本课程采用了理论和实践一体化并结合交互式的教学方法。本文以基于深度学习的图像分割算法为例，讲解理论知识和对应的实例，采用交互式方法让学生参与到实验过程中。在本例中当学生选择不同的交互点时会产生不同的图像分割结果。这种教学方法将可以让学生对所学知识更加融会贯通。

关键词

计算机视觉；深度学习；交互式；图像分割

参考

[1]高常鑫,桑农,王岳环.大学本科开设计算机视觉课程教学的探讨[J].湖北第二师范学院学报,2015,32(08):96-99.

[2]朱虎,卢官明,谢世朋. “数字电视原理”理论实践一体化教学研究[J].电子世界,2014(13):54.

[3]卢旭,刘钊.基于深度学习的图像语义分割技术综述[J].软件导刊,2021,20(01):242-244.

[4]Minaee S, Boykov YY, Porikli F, et al. Image seg-mentation using deep learning: A survey[J]. IEEE Transac-tions on Pattern Analysis and Machine Intelligence, 2021.

[5]Harikrishnan J, Sudarsan A, Sadashiv A, et al.Vision-face recognition attendance monitoring system forsurveillance using deep learning technology and computervision[C]//2019 International Conference on Vision To-wards Emerging Trends in Communication and Networking(ViTECoN). IEEE, 2019: 1-5.

[6]Dickmanns E D, Zapp A. Autonomous high speed roadvehicle guidance by computer vision[J]. IFAC ProceedingsVolumes, 1987, 20(5): 221-226.

[7]Wang G, Zuluaga M A, Li W, et al. DeepIGeoS: a deepinteractive geodesic framework for medical imagesegmentation[J]. IEEE transactions on pattern analysisand machine intelligence, 2018, 41(7): 1559-1572.

[8]Qu Z, Mei J, Liu L, et al. Crack detection ofconcrete pavement with cross-entropy loss function andimproved VGG16 network model[J]. IEEE Access, 2020, 8:54564-54573.

[9]Klein S, Pluim J P W, Staring M, et al. Adaptivestochastic gradient descent optimisation for imageregistration[J]. International journal of computer vision,2009, 81(3): 227.

[10]Tovar E, Vasques F, Burns A. Supporting real-timedistributed computer-controlled systems with multi-hop P-NET networks[J]. Control Engineering Practice, 1999, 7(8): 1015-1025.

[11]Nobis F, Geisslinger M, Weber M, et al. A deeplearning-based radar and camera sensor fusion architec-ture for object detection[C]//2019 Sensor Data Fusion:Trends, Solutions, Applications (SDF). IEEE, 2019: 1-7.

[12]Viprey V F, Corrias M V, Kagedal B, et al.Standardisation of operating procedures for the detectionof minimal disease by QRT-PCR in children with neuroblastoma:quality assurance on behalf of SIOPEN-R-NET[J]. EuropeanJournal of Cancer, 2007, 43(2): 341-350.