Deep Fake Detection using machine learning.pptx

It is a presentation about deepfake detection

1. Deep Fake Detection Using Machine Learning ~By: Mr. Ahmed Alaini , Mr. Shehab Alkehtani , Mr. Bessam Alarifi , Mr. Abdulwahab Alsamawai Guided By : Prof. Malek Algabri Sana’a University Faculty of Computer and Technology August 2024

2. Content Introduction Literature Review Research objectives Research problems Significance and expected contribution Research Methodology Timeline Conclusion

3. What is Deep Fake?

4. Introduction Deep fake refers to synthetically generated content, primarily fake images and videos, created using AI which appear to be real. Potential Benefits of Deep fake Technology. Potential Risks of Deep fake Technology. The Growing Problem of Deep fake. Need for Solutions.

5. Conclusion Publication Year Author Paper Name Sr No This survey provides a timely overview of deepfake creation and detection methods. 2019 Thanh Thi Nguyen, Cuong M. Ngunyen Deep Learning for DeepFakes Creation and Detection 1 In this paper, we have presented a brief review of some papers which describes different methods to detect deepfake videos and images. 2021 Karthik P C, Sanjana Review of DeepFake Detection Techniques 2 Our comprehensive experiments demonstrated the feasibility of building a general detection method to deal with manipulated images and videos. 2018 Huy H. Nguyen, Junichi Yamagishi , and Isao Echizen CAPSULE-FORENSICS: USING CAPSULE NETWORKS TO DETECT FORGED IMAGES AND VIDEOS 3 In this paper, we present FakeCatcher , a fake portrait video detector based on biological signals. 2020 Umur Aybars Ciftci , Ilke Demir , and Lijun Yin FakeCatcher : Detection of Synthetic Portrait Videos using Biological Signals 4 Literature Survey

6. Existing Gaps: Most existing models are trained on specific datasets and may not generalize well to different real-world scenarios. Deep fake detection models are often susceptible to adversarial attacks, which can manipulate the model's predictions. Making it difficult to understand their decision-making process

7. Research objectives 1- Overall Objective 2- Specific Objectives

8. Research objectives Overall Objective To develop and evaluate machine learning models for effective and accurate detection of deep fake video and images across real-word scinarios .

9. Research objectives Specific Objectives Ethical Considerations Dataset Exploration improve generalizability. Model Development Model Evaluation

10. Research problems Technical Challenges Data-Related Challenges Deployment and Integration Challenges Communicating Difficulties in the group of researchers

11. Significance and expected contributions 1- Significance : Growing Social Issue Law Enforcement Enhance Social Media Platforms Guiding Future Research

12. Significance and Expected Contributions 2- Expected contributions: Critical Evaluation of Deep Learning High-Performance Deep fake detection model Emphasis on Data Quality Identification of Research Gaps  

13. Research Methodology Data Collection and Preprocessing Model Development Model Evaluation Adversarial Testing

14. Data Collection and Preprocessing Gather a comprehensive dataset of real and deep fake media, ensuring diversity in content, generation techniques, and quality.

15. Model Development  Design and train a deep learning model using convolutional neural networks (CNNs), recurrent neural networks (RNNs), or other suitable architectures.

16. Model Evaluation Evaluate the model's performance on diverse datasets, including benchmark datasets and real-world examples.

17. Adversarial Testing Investigate the model's resistance to adversarial attacks and different deep fake generation techniques.

18. Timeline Weak 1 :  Dataset collection, preprocessing, and model selection. Weak 2 :  Model training and optimization, initial evaluation. Weak 3 :  Adversarial testing. Weak 4 :  Final evaluation, report writing, and dissemination of findings.