Spatiotemporal Trident Networks: Detection and Localization of Object Removal Tampering in Video Passive Forensics

Yang, Quanxin; Yu, Dongjin; Zhang, Zhuxi; Yao, Ye; Chen, Linqiang

doi:10.1109/tcsvt.2020.3046240

Cited by 26 publications

(18 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Adding or removing objects from a certain frame can cause an abrupt change in a video sequence. Because the motion residual can act as a clue in the detection of the object-based frame tampering, it is frequently used in detecting video tampering [17,[23][24][25]. However, a scene change in a video frame can also cause an abrupt change.…”

Section: Motion Residualmentioning

confidence: 99%

“…From Table 2, we can see that both the non-forged and tampered-with frames are identified well with an accuracy of 98% or more. We compared our with five state-of-the-art methods by Yao et al [23] (CNN), Kohli et al [24] (Temporal CNN), Aloraini et al [31] (TF-SA), Aloraini et al [21] (S-PA), and Yang et al [25] (STN). Table 3 shows six metrics for six forgery detection algorithms, including the proposed method.…”

Section: Two-class Identificationmentioning

confidence: 99%

“…They employed the motion residual to train the presented network. Yang et al [25] proposed a spatiotemporal trident network based on the spatial rich model and three-dimensional convolution, which provides branches and can theoretically improve the detection and localization accuracy of tampered regions.…”

Section: Introductionmentioning

confidence: 99%

“…Motion residual was widely used as a key feature to detect object-based video forgery [17,[23][24][25]. For each video frame, the motion residual contains a substantial number of the intra-frame properties of that frame as well as the inter-frame inherent properties of the corresponding frame.…”

Section: Introductionmentioning

confidence: 99%

“…Most past studies [21,23,25] performed two-class identification, which classifies whether the video frame has been tampered with, to detect object-based video forgery. In [24], another two-class identification is defined by identifying whether the frame has been forged or double-compressed.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Frame Identification of Object-Based Video Tampering Using Symmetrically Overlapped Motion Residual

2022

View full text Add to dashboard Cite

Image and video manipulation has been actively used in recent years with the development of multimedia editing technologies. However, object-based video tampering, which adds or removes objects within a video frame, is posing challenges because it is difficult to verify the authenticity of videos. In this paper, we present a novel object-based frame identification network. The proposed method uses symmetrically overlapped motion residuals to enhance the discernment of video frames. Since the proposed motion residual features are generated on the basis of overlapped temporal windows, temporal variations in the video sequence can be exploited in the deep neural network. In addition, this paper introduces an asymmetric network structure for training and testing a single basic convolutional neural network. In the training process, two networks with an identical structure are used, each of which has a different input pair. In the testing step, two types of testing methods corresponding to two- and three-class frame identifications are proposed. We compare the identification accuracy of the proposed method with that of the existing methods. The experimental results demonstrate that the proposed method generates reasonable identification results for both two- and three-class forged frame identifications.

show abstract

Section: Motion Residualmentioning

confidence: 99%