Deepfake Detection Using Artificial Intelligence and Convolutional Neural Networks.

This research is about a way to detect fake videos using a special kind of computer program called a Convolutional Neural Network. This program looks at the faces in the videos to see if they are real or not. It does this by breaking down the video into parts finding the faces and then looking at the faces in detail. The program is different from ways of detecting fake videos because it looks at different parts of the face like the eyes and nose and not just the whole face. This helps it find clues that might show the video is fake. The program was tested on some videos that are known to be fake. It was very good at telling which ones were real and which ones were not. It was also very good at not making mistakes and saying a real video was fake. There are some challenges when trying to detect videos like if the video is not very clear or if it is compressed.. Even with these challenges the program was still very good at detecting fake videos. It is very important to be able to detect videos because they can be used to trick people on the internet in politics and in other places. This research helps to make the internet a safer place by making it harder for people to spread videos. Overall this program is a way to detect fake videos and can be used as a starting point for more research, in this area. The Convolutional Neural Network-based deepfake detection framework is a way to identify manipulated videos and it can help to prevent the spread of misinformation online. The deepfake detection system can be used to make sure that videos are real and it can help to keep people from fake videos.

[1] A. Rossler and his team wrote a paper called "FaceForensics++: Learning to Detect Manipulated Facial Images" for the IEEE International Conference on Computer Vision in 2019. They focused on face manipulation detection. [2] F. Chollet wrote about "Xception: Deep Learning with Depthwise Separable Convolutions" in the Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition in 2017. This paper discussed Xception. [3] D. Wodajo and S. Atnafu made a report called "Deepfake Video Detection Using Convolutional Vision Transformer". They posted it on arXiv in 2021. The report is about deepfake detection. [4] Y. Li, M. Chang and S. Lyu did a study called "In Ictu Oculi: Exposing AI Generated Videos by Detecting Eye Blinking" for the IEEE International Workshop on Information Forensics and Security in 2018. The study looked at AI generated videos. [5] N. Sharma and R. K. Sharma wrote an article called "Detecting Deepfakes Using Convolutional Neural Networks" for the International Journal of Computer Techniques. It was published in volume 12 number 3 pages 1–4 in 2025. Deepfakes were the topic. [6] A. H. Soudy and his team wrote about "Deepfake Detection Using Convolutional Vision Transformers and Convolutional Neural Networks" in Neural Computing and Applications. The article was published in volume 36 pages 19759–19775 in 2024. Deepfake detection was discussed. [7] Facebook AI has a dataset called "DeepFake Detection Challenge Dataset (DFDC)" which's available at https://ai.facebook.com/datasets/dfdc. This dataset helps with deepfake detection. [8] A. Dosovitskiy and other people made a report. They called it "An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale". They put this report on arXiv in the year 2020. The report is about image recognition. A. Dosovitskiy and other people worked on this report. [9] Y. Li and other people created a dataset. The dataset is called "Celeb-DF: A Large-Scale Challenging Dataset for DeepFake Forensics". They made this dataset for the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Y. Li and other people presented this dataset in the year 2020. The dataset is, for DeepFake forensics. Y. Li and other people worked on the Celeb-DF dataset. [10] T. Zhang wrote a survey called "Deepfake Generation and Detection: A Survey”, for Multimedia Tools and Applications. The survey was published in volume 81 number 5 pages 6259–6276 in 2022. The survey covered deepfake generation and detection.