Skip to content

Opening book details…

Can I read Video Forensics for Object Removal Based on Darknet3D on EtoBox?

Video Forensics for Object Removal Based on Darknet3D by Kejun Zhang; Yuhao Wang; Xinying Yu is a book available to read on EtoBox.

What is Video Forensics for Object Removal Based on Darknet3D about?

To address the problems of insufficient analysis of timedomain information of tampered video by 2D convolutional neural networks and the loss of details in the pooling layer when processing frame images, a 3D video object removal tamper detection and localization model based on Darknet53 optimization network is proposed. While the Darknet53 network can fully retain the detail information in the frame, we try to give the two-dimensional Darknet53 network, which can only process spatial information, the ability to process time-domain information, and extend the two-dimensional convolutional layer into a threedimensional convolutional layer, and also improve the detection efficiency by adjusting the network structure to reduce feature redundancy, making it more suitable for efficient processing of video tampering detection binary classification tasks. A D3D (Darknet3D) network is constructed to improve feature adequacy representation. Experimental results reveal that the temporal domain classification accuracy of the tamper detection model based on the Darknet3D is 98.9%, and the average Intersection over Union of spatial localization and tamper area labeling is 49.7%, which can effec

Author
Kejun Zhang; Yuhao Wang; Xinying Yu
Publisher
Springer International Publishing Springer
Published
2022
Language
EN
ISBN
9783031157776
Subjects
Business, Computer Science, Management

More by Kejun Zhang; Yuhao Wang; Xinying Yu

Browse all works by Kejun Zhang; Yuhao Wang; Xinying Yu

Similar books