当前位置: X-MOL 学术J. Real-Time Image Proc. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
The role of film and television big data in real-time image detection and processing in the Internet of Things era
Journal of Real-Time Image Processing ( IF 2.9 ) Pub Date : 2021-04-27 , DOI: 10.1007/s11554-021-01105-y
Yangfan Tong , Wei Sun

With the rapid development of the Internet, images play an increasingly critical role as an important data source in the Internet of Things (IoT). To meet the low-latency and high-efficiency transmission method of the IoT platform, the current problems in the image processing are analyzed from the perspective of real-time image processing in this study. The image features are extracted with the back propagation neural network (BPNN), and the images are classified with the support vector machines (SVM). A real-time image detection and processing platform (RT-IDPP) is constructed using the Adaboost framework based on the IoT, and the real-time image transmission and processing is realized based on different databases. It is found that the RT-IDPP proposed for the IoT realizes the image detection and tracking. The proposed method can not only run effectively on different cloud platforms for use, but also meet the real-time requirements in the image detection and tracking process, ensuring that the image detection rate is higher than 97%. Thus, the detection effect is better. Compared with the traditional image detection methods, the proposed method has higher detection rate and lower false-negative rate (FNR) and false-positive rate (FPR). The experimental detection effect on the film and television big data (FTBD) database is significantly better than that of other databases. This research can provide a theoretical basis for related researches on real-time image processing in the environment of IoT.



中文翻译:

影视大数据在物联网时代实时图像检测与处理中的作用

随着Internet的快速发展,图像作为物联网(IoT)中的重要数据源,发挥着越来越重要的作用。为了满足物联网平台的低延迟,高效率的传输方式,本文从实时图像处理的角度分析了当前图像处理中存在的问题。利用反向传播神经网络(BPNN)提取图像特征,并利用支持向量机(SVM)对图像进行分类。利用基于物联网的Adaboost框架构建了实时图像检测与处理平台(RT-IDPP),并基于不同的数据库实现了实时图像的传输与处理。结果发现,针对物联网提出的RT-IDPP实现了图像检测和跟踪。该方法不仅可以在不同的云平台上有效运行,而且可以满足图像检测和跟踪过程中的实时性要求,确保图像检测率高于97%。因此,检测效果更好。与传统的图像检测方法相比,该方法具有较高的检测率和较低的假阴性率(FNR)和假阳性率(FPR)。影视大数据(FTBD)数据库的实验检测效果明显优于其他数据库。该研究可以为物联网环境下实时图像处理的相关研究提供理论依据。确保图像检测率高于97%。因此,检测效果更好。与传统的图像检测方法相比,该方法具有较高的检测率和较低的假阴性率(FNR)和假阳性率(FPR)。影视大数据(FTBD)数据库的实验检测效果明显优于其他数据库。该研究可以为物联网环境下实时图像处理的相关研究提供理论依据。确保图像检测率高于97%。因此,检测效果更好。与传统的图像检测方法相比,该方法具有较高的检测率和较低的假阴性率(FNR)和假阳性率(FPR)。影视大数据(FTBD)数据库的实验检测效果明显优于其他数据库。该研究可以为物联网环境下实时图像处理的相关研究提供理论依据。影视大数据(FTBD)数据库的实验检测效果明显优于其他数据库。该研究可以为物联网环境下实时图像处理的相关研究提供理论依据。影视大数据(FTBD)数据库的实验检测效果明显优于其他数据库。该研究可以为物联网环境下实时图像处理的相关研究提供理论依据。

更新日期:2021-04-28
down
wechat
bug