Convolutional neural network initialization approaches for image manipulation detection

Camacho, Ivan Castillo; Wang, Kai

doi:10.1016/j.dsp.2021.103376

Cited by 5 publications

(3 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Because only one branch participates in coaching, the number of parameters decreases; therefore, the range of layers of the network is additionally reduced, leading to an associated degree of improvement within the speed of forward and backward propagation of the network. Usually, we train 50-100 epochs for the initialization of the network [2,3]. We used the JTA and Microsoft COCO 2014 datasets for joint coaching (Figure 4).…”

Section: Training Experimentsmentioning

confidence: 99%

“…The keys to pose recognition and two-dimensional (2D) animation generation of characters lie in pose extraction, image compression, and the subsequent extraction and application of each action of the character itself. Convolutional neural networks (CNN) have shown great potential for medical image processing [3,4].…”

Section: Introductionmentioning

confidence: 99%

“…However, this method is more challenging and the system complexity is higher. (3) The researcher attempts to generate real-time character animation that conforms to the characteristics of the surrounding environment and character behavior based on the collected 3D posture information, using the surrounding environment and character behavior as input. This method can significantly reduce the workload of animation creators and simplify the complexity of creating high-fidelity virtual character animations.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

High Speed and Accuracy of Animation 3D Pose Recognition Based on an Improved Deep Convolution Neural Network

Ding

2023

Applied Sciences

View full text Add to dashboard Cite

Pose recognition in character animations is an important avenue of research in computer graphics. However, the current use of traditional artificial intelligence algorithms to recognize animation gestures faces hurdles such as low accuracy and speed. Therefore, to overcome the above problems, this paper proposes a real-time 3D pose recognition system, which includes both facial and body poses, based on deep convolutional neural networks and further designs a single-purpose 3D pose estimation system. First, we transformed the human pose extracted from the input image to an abstract pose data structure. Subsequently, we generated the required character animation at runtime based on the transformed dataset. This challenges the conventional concept of monocular 3D pose estimation, which is extremely difficult to achieve. It can also achieve real-time running speed at a resolution of 384 fps. The proposed method was used to identify multiple-character animation using multiple datasets (Microsoft COCO 2014, CMU Panoptic, Human3.6M, and JTA). The results indicated that the improved algorithm improved the recognition accuracy and performance by approximately 3.5% and 8–10 times, respectively, which is significantly superior to other classic algorithms. Furthermore, we tested the proposed system on multiple pose-recognition datasets. The 3D attitude estimation system speed can reach 24 fps with an error of 100 mm, which is considerably less than that of the 2D attitude estimation system with a speed of 60 fps. The pose recognition based on deep learning proposed in this study yielded surprisingly superior performance, proving that the use of deep-learning technology for image recognition has great potential.

show abstract

Section: Training Experimentsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

High Speed and Accuracy of Animation 3D Pose Recognition Based on an Improved Deep Convolution Neural Network

Ding

2023

Applied Sciences

View full text Add to dashboard Cite

show abstract

Providing robust and reversible watermarking algorithm and detection of malicious manipulation

Damghani

Estaji

Damghani

2022

Security and Privacy

View full text Add to dashboard Cite

In this article, due to the importance of embedding encrypted information in the image, a watermarking method is proposed based on local statistics in pixels. The advantages of the proposed method include its complete reversibility and very high security, which is due to the use of the quadtree concept, finding similar parts in the image, and embedding information in low‐energy and less significant pixels. This method is an alternative and simplifier of the quadtree method, dividing the image into identical non‐overlapping blocks and then choosing the block with less energy as more suitable for watermarking. Also, two criteria of mean‐squared‐error and peak signal‐to‐noise ratio (PSNR) have been calculated and the PSNR shows that image quality has not changed too much after hiding the information than the original image. It should be noted that since the proposed method is applied to black and white images, the result can be generalized to different color channels in the image, and color images have more visible capabilities in terms of cached memory than black and white images because they contain three dimensions of information.

show abstract

Feature enhancement and supervised contrastive learning for image splicing forgery detection

Zheng

Fang

et al. 2023

Digital Signal Processing

View full text Add to dashboard Cite

Convolutional neural network initialization approaches for image manipulation detection

Cited by 5 publications

References 40 publications

High Speed and Accuracy of Animation 3D Pose Recognition Based on an Improved Deep Convolution Neural Network

High Speed and Accuracy of Animation 3D Pose Recognition Based on an Improved Deep Convolution Neural Network

Providing robust and reversible watermarking algorithm and detection of malicious manipulation

Feature enhancement and supervised contrastive learning for image splicing forgery detection

Contact Info

Product

Resources

About