deepfake seminar computer engineering ppt on deepfake which uses ai and deep learning technology.with adavantages,disadvantages,intro,reference,conclusion

1. Seminar and Technical Communication
DEEP FAKE

2. Under the guidance of
Prof. Swamiraj Jadhav
Chinchole Suraj Pramod
20CO018

3. How DeepFake
is Created ?
Advantages and
Disadvantages
Introduction
01 02
TABLE OF CONTENTS
03
References
Conclusion
04 05

4.  Deep fake (also spelled deepfake) is a type of
artificial intelligence used to create
convincing images, audio and video hoaxes.
 Deepfake is a technique that uses deep
learning algorithms to create fake images
usually by swapping a person’s face from a
source image into another person’s face in a
target image, with a resulting fake image that
is sometimes hard to detect.
INTRODUCTION

5. How DeepFake is created ?
 Deepfakes are created using deep learning methods in
which they aim to replace the face of a targeted
person by the face of someone else in an image or
video.
 This technique was improved by developers and
online communities to notably create user-friendly
applications that are readily available online like
FakeApp and FaceSwap.
 Deepfake videos are often made using a variational
auto-encoder (VAE) and a facial recognition algorithm.
 Images are encoded into low-dimensional
representations, which are then decoded back into
images by trained VAEs.

6. Advantages of DeepFake
 Low-Cost Video Campaigns - Marketers using
deepfakes can save money on the budgets for
their video campaigns using because you don’t
need an in-person actor.
 Improved Omni Channel Campaigns-
Instead of reshooting to fit different purposes for
different channels, you can instead just edit or
replace video cuts to create a paid social
campaign.
 Hyper Personalization- Deepfake technology
would allow you to alter the skin tone of that
model so the customer can experience what the
product would look like on their skin tone.

7. Disadvantages of DeepFake
 Scamming Increases - Deepfake technology
may also increase the number of scams online,
you could create false accusations or complaints
against companies.
 Lack Of Trust Or Ethics Issues - The most
obvious impact of deepfake technology is that it
can be used to create a fake video, so ascertaining
the authenticity of a piece of content has become
more difficult.
 Spreading Misleading News via Politicians -
We are so open to believing what we see or hear
in media. And if you see a public figure talking
about a topic, you usually would not think if the
person is “real” or “fake?”.

8. DeepFake Detection
 Deepfake Detection To successfully build a robust
neural network capable of detecting complex
deepfakes a large dataset of training images is needed.
due to social media these kind of images can be easily
obtained also companies like google provide large
datasets to help accelerate the research in protection
against deepfakes . MesoNet used a dataset of more
than 5000 images Fig. 6, the images are divided into
real images and deepfake images.
 After training the CNN with the dataset, the CNN is able
to detect deepfake images with over 80 % confidence
rate Figure. When exploring current deepfake images
available online, deepfake flaws are easily noticeable,
deepfake images of a person’s face looking straight
ahead to the camera are harder to detect than images
where the person’s face is looking at an angle. current
deepfake generation techniques aren’t that great at
dealing with faces at an angle.

9. Conclusion
• People got used to hearing about Deepfakes.
Most of the Internet people know what
Deepfake is. We do not have to wait for years
to reach and use new technologies like old
times.
• We are in the era of Artificial Intelligence. They
can see, they can analyze, and they can fake it.
There are good and bad sides of Deepfakes,
but the usage area of technology depends on
us. We must use it for the greater good. Also,
some legal arrangements must be done.
• Indeed, all these will take some time to reach
a certain point. We will see what kind of world
this technology leads us to.

10. References
1. Deepfakes Creation and Detection Using Deep Learning Hady A.
Khalil, Shady A. Maged Department o f Mechatronics
Engineering, Ain Shams University, Cairo, Egypt
hadyayman1996@gmail.com, shady.maged@eng.asu.edu.eg
(https://ieeexplore.ieee.org/document/9447642)
2. Deepfake Detection through Deep Learning Deng Pan, Lixian
Sun, Rui Wang, Xingjian Zhang, Richard O. Sinnott School of
Computing and Information Systems The University of Melbourne,
Melbourne, Australia Contact: rsinnott@unimelb.edu.au
(https://ieeexplore.ieee.org/document/9302547)

11. Thank You..