Scribd
Upload
15 views4 pages

Atique

Uploaded by

alamahmedraza
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
15 views4 pages

Atique

Uploaded by

alamahmedraza
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

A PROJECT REPORT

On

Deep fake Video Detection


Submitted by

Mr Atique Syed
in partial fulfilment for the award of the degree
of

BACHELOR OF SCIENCE
in
COMPUTER SCIENCE

under the guidance of

Ms. Riddhi Pandya


Department of Computer Science

Thakur Ramnarayan College Of Arts And Commerce


(Sem V)
(2024 – 2025)
Thakur Ramnarayan College of Arts and Commerce
Thakur Ramnarayan Educational Campus, S.V. Road, Dahisar (East), Mumbai-400068

Department of Computer Science

CERTIFICATE

This is to certify that Mr. Atique syed of T.Y.B.Sc.(Computer Science)(Sem V)


class has satisfactorily completed the Project Deepfake Video Detection, to be submitted in
the partial fulfilment for the award of Bachelor of Science in Computer Science during the
academic year 2024 – 2025

Date of Submission:

Project Guide Head / Incharge


Department Computer Science

College Seal Signature of examiner


1.Title : DeepFake Video Detection

2. Introduction:
With the rise of AI-generated deep fake videos, distinguishing real videos from fake ones has
become a pressing issue. Deep Fakes manipulate video content to create highly realistic fake
videos that are challenging to detect. This project aims to develop a machine learning-based
system to identify deepfake videos by analysing subtle inconsistencies in facial features,
movements, and textures.By focusing on identifying artifacts within frames, inconsistencies
in facial movements, and pixel-level distortions, this study proposes a method that enhances
forensic investigation capabilities and prevents misuse of video-based content. Furthermore,
this study explores the future possibilities for improving detection accuracy by integrating
more advanced neural network architectures.

3. Objectives:

● To develop a machine learning system capable of detecting deepfake videos with high
accuracy.
● To analyse video frames and extract key features that differentiate real videos from
deep fakes.
● To create an easy-to-use interface where users can upload and classify videos as
"FAKE" or "REAL."

4. Scope:

● This project will focus on detecting deep fake videos generated using popular
methods, including GAN-based techniques. It can be applied in social media
platforms, media verification, and digital security to combat misinformation and
fraud.

5. Methodology:

● Data Collection: Use deepface video datasets such as the FaceForensics++ dataset.
● Preprocessing: Extract frames from videos and apply facial landmark detection.
● Feature Extraction: Use deep learning models (e.g., CNNs) to extract important
features from the video frames.
● Model Training: Train a deep learning model to classify videos as either real or fake
based on extracted features.
● Testing & Evaluation: Split the dataset into training and testing sets to evaluate the
model’s performance.
● Deployment: Build a user-friendly interface for video upload and classification.
6. Tools and Technologies:

● Programming Language: Python


● Libraries: OpenCV, TensorFlow, Keras, dlib, matplotlib
● Datasets: FaceForensics++, Deepfake Detection Challenge dataset from Kaggle
● Development Environment: Visual Studio Code, Anaconda for package management

7. Timeline:

● Week 1: Project planning and dataset collection


● Week 2-3: Frame extraction and feature analysis
● Week 4-5: Model training and testing
● Week 6: Model fine-tuning and evaluation
● Week 7: Deployment and project documentation

8. Resources:

● Hardware: Computer with sufficient GPU for deep learning model training
● Software: Python, deep learning frameworks, cloud storage for large datasets
● Human Resources: Project guidance from supervisor and relevant research papers for
reference

9. Expected Outcomes:

● A machine learning model capable of detecting deep fake videos with high accuracy.
● A user interface for uploading videos and receiving classification results as either
"FAKE" or "REAL."
● Contributions to combating deepfake misinformation in media and security fields.

10. References:
1. Rossler, A., Cozzolino, D., Verdoliva, L., et al. "FaceForensics++: Learning to Detect
Manipulated Facial Images." ICCV, 2019.
2. Karras, T., Laine, S., Aila, T. "A Style-Based Generator Architecture for Generative
Adversarial Networks." CVPR, 2019.

You might also like