Deep fake Detection Through Deep Learning
2
Citation
4
Reference
10
Related Paper
Citation Trend
Abstract:
Abstract: Deep fake is a rapidly growing concern in society, and it has become a significant challenge to detect such manipulated media. Deep fake detection involves identifying whether a media file is authentic or generated using deep learning algorithms. In this project, we propose a deep learning-based approach for detecting deep fakes in videos. We use the Deep fake Detection Challenge dataset, which consists of real and Deep fake videos, to train and evaluate our deep learning model. We employ a Convolutional Neural Network (CNN) architecture for our implementation, which has shown great potential in previous studies. We pre-process the dataset using several techniques such as resizing, normalization, and data augmentation to enhance the quality of the input data. Our proposed model achieves high detection accuracy of 97.5% on the Deep fake Detection Challenge dataset, demonstrating the effectiveness of the proposed approach for deep fake detection. Our approach has the potential to be used in real-world scenarios to detect deep fakes, helping to mitigate the risks posed by deep fakes to individuals and society. The proposed methodology can also be extended to detect in other types of media, such as images and audio, providing a comprehensive solution for deep fake detection.Keywords:
Deep Neural Networks
Normalization
Deep learning has been very successful in dealing with big data from various fields of science and engineering. It has brought breakthroughs using various deep neural network architectures and structures according to different learning tasks. An important family of deep neural networks are deep convolutional neural networks. We give a survey for deep convolutional neural networks induced by 1‐D or 2‐D convolutions. We demonstrate how these networks are derived from convolutional structures, and how they can be used to approximate functions efficiently. In particular, we illustrate with explicit rates of approximation that in general deep convolutional neural networks perform at least as well as fully connected shallow networks, and they can outperform fully connected shallow networks in approximating radial functions when the dimension of data is large.
Deep Neural Networks
Cite
Citations (9)
Fire detection using computer vision techniques and image processing has been a topic of interest among the researchers. Indeed, good accuracy of computer vision techniques can outperform traditional models of fire detection. However, with the current advancement of the technologies, such models of computer vision techniques are being replaced by deep learning models such as Convolutional Neural Networks (CNN). However, many of the existing research has only been assessed on balanced datasets, which can lead to the unsatisfied results and mislead real-world performance as fire is a rare and abnormal real-life event. Also, the result of traditional CNN shows that its performance is very low, when evaluated on imbalanced datasets. Therefore, this paper proposes use of transfer learning that is based on deep CNN approach to detect fire. It uses pre-trained deep CNN architecture namely VGG, and MobileNet for development of fire detection system. These deep CNN models are tested on imbalanced datasets to imitate real world scenarios. The results of deep CNNs models show that these models increase accuracy significantly and it is observed that deep CNNs models are completely outperforming traditional Convolutional Neural Networks model. The accuracy of MobileNet is roughly the same as VGGNet, however, MobileNet is smaller in size and faster than VGG.
Transfer of learning
Deep Neural Networks
Cite
Citations (55)
Deep Learning has achieved tremendous results by pushing the frontier of automation in diverse domains. Unfortunately, current neural network architectures are not explainable by design. In this paper, we propose a novel method that trains deep hypernetworks to generate explainable linear models. Our models retain the accuracy of black-box deep networks while offering free lunch explainability by design. Specifically, our explainable approach requires the same runtime and memory resources as black-box deep models, ensuring practical feasibility. Through extensive experiments, we demonstrate that our explainable deep networks are as accurate as state-of-the-art classifiers on tabular data. On the other hand, we showcase the interpretability of our method on a recent benchmark by empirically comparing prediction explainers. The experimental results reveal that our models are not only as accurate as their black-box deep-learning counterparts but also as interpretable as state-of-the-art explanation techniques.
Interpretability
Black box
Benchmark (surveying)
Deep Neural Networks
Cite
Citations (1)
Cite
Citations (5)
Purpose The paper aims to address the tracking algorithm based on deep learning and four deep learning tracking models developed. They compared with each other to prevent collision and to obtain target tracking in autonomous aircraft. Design/methodology/approach First, to follow the visual target, the detection methods were used and then the tracking methods were examined. Here, four models (deep convolutional neural networks (DCNN), deep convolutional neural networks with fine-tuning (DCNNFN), transfer learning with deep convolutional neural network (TLDCNN) and fine-tuning deep convolutional neural network with transfer learning (FNDCNNTL)) were developed. Findings The training time of DCNN took 9 min 33 s, while the accuracy percentage was calculated as 84%. In DCNNFN, the training time of the network was calculated as 4 min 26 s and the accuracy percentage was 91%. The training of TLDCNN) took 34 min and 49 s and the accuracy percentage was calculated as 95%. With FNDCNNTL, the training time of the network was calculated as 34 min 33 s and the accuracy percentage was nearly 100%. Originality/value Compared to the results in the literature ranging from 89.4% to 95.6%, using FNDCNNTL, better results were found in the paper.
Transfer of learning
Tracking (education)
Cite
Citations (16)
The suggested study's objectives are to develop an unique criterion-based method for classifying RBC pictures and to increase classification accuracy by utilizing Deep Convolutional Neural Networks instead of Conventional CNN Algorithm. Materials and Procedures A dataset-master image dataset of 790 pictures is used to apply Deep Convolutional Neural Network. Convolutional Neural Network and Deep Convolutional Neural Network comparison using deep learning has been suggested and developed to improve classification accuracy of RBC pictures. Using Gpower, the sample size was calculated to be 27 for each group. Results: When compared to Convolutional Neural Network, Deep Convolutional Neural Network had the highest accuracy in classifying blood cell pictures (95.2%) and the lowest mean error (85.8 percent). Between the classifiers, there is a statistically significant difference of p=0.005. The study demonstrates that Deep Convolutional Neural Networks perform more accurately than Conventional Neural Networks while classifying photos of blood cells[1].
Convolution (computer science)
Cite
Citations (2)
Batch Normalization (BN) is widely used in {centralized} deep learning to improve convergence and generalization. However, in {federated} learning (FL) with decentralized data, prior work has observed that training with BN could hinder performance and suggested replacing it with Group Normalization (GN). In this paper, we revisit this substitution by expanding the empirical study conducted in prior work. Surprisingly, we find that BN outperforms GN in many FL settings. The exceptions are high-frequency communication and extreme non-IID regimes. We reinvestigate factors that are believed to cause this problem, including the mismatch of BN statistics across clients and the deviation of gradients during local training. We empirically identify a simple practice that could reduce the impacts of these factors while maintaining the strength of BN. Our approach, which we named FIXBN, is fairly easy to implement, without any additional training or communication costs, and performs favorably across a wide range of FL settings. We hope that our study could serve as a valuable reference for future practical usage and theoretical analysis in FL.
Normalization
Cite
Citations (2)
On our planet, skin cancer is among the most dangerous diseases. It is, however, difficult to diagnose skin cancer correctly. A variety of tasks have recently been shown to be excelled by machine learning and deep learning algorithms. In the case of skin diseases, these algorithms are very useful. In this article, we examine various machine learning and deep learning techniques and their use in diagnosing skin diseases. In this paper, we discuss common skin diseases and the method of acquiring images from dermatology, and we present several freely available datasets. Our focus shifts to exploring popular machine learning and deep learning architectures and popular frameworks for implementing machine and deep learning algorithms once we have introduced machine learning and deep learning concepts. Following that, performance evaluation metrics are presented. Here we are going to review the literature on machine and deep learning and how these technologies can be used to detect skin diseases. Furthermore, we discuss potential research directions and the challenges in the area. In this paper, the principal goal is to describe contemporary machine learning and deep learning methods for skin disease diagnosis
Cite
Citations (1)
Deep learning is now an active research area. Deep learning has done a success in computer vision and image recognition. It is a subset of the Machine Learning. In Deep learning, Convolutional Neural Network (CNN) is popular deep neural network approach. In this paper, we have addressed that how to extract useful leaf features automatically from the leaf dataset through Convolutional Neural Networks (CNN) using Deep Learning. In this paper, we have shown that the accuracy obtained by CNN approach is efficient when compared to accuracy obtained by the traditional neural network.
Deep Neural Networks
Cite
Citations (7)
Differentially private stochastic gradient descent (DPSGD) is a variation of stochastic gradient descent based on the Differential Privacy (DP) paradigm, which can mitigate privacy threats that arise from the presence of sensitive information in training data. However, one major drawback of training deep neural networks with DPSGD is a reduction in the models accuracy. In this paper, we study the effect of normalization layers on the performance of DPSGD. We demonstrate that normalization layers significantly impact the utility of deep neural networks with noisy parameters and should be considered essential ingredients of training with DPSGD. In particular, we propose a novel method for integrating batch normalization with DPSGD without incurring an additional privacy loss. With our approach, we are able to train deeper networks and achieve a better utility-privacy trade-off.
Normalization
Stochastic Gradient Descent
Deep Neural Networks
Differential Privacy
Database normalization
Training set
Cite
Citations (4)