International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 04 | Apr 2025
p-ISSN: 2395-0072
www.irjet.net
DEEPFAKE DETECTION USING MACHINE LEARNING Chandan Mahto1, Prapti Khaparde2, Priya Khatake3 , Dr. Rajesh Kadu4 1,2,3Student at Mahatma Gandhi Mission’s College of Engineering And Technology, Navi Mumbai 4Associate Professor at Mahatma Gandhi Mission’s College of Engineering And Technology, Navi Mumbai
---------------------------------------------------------------------***--------------------------------------------------------------------Abstract – With the increasing misuse of artificial 2. Motivation intelligence for generating deepfakes in the form of fake images, cloned voices, and manipulated videos, ensuring media authenticity has become a significant challenge. This paper presents a unified machine learning-based multi-modal deepfake detection system capable of detecting forgeries in audio, image, and video formats. For image-based detection, a pipeline using MTCNN for face detection and InceptionResNetV1 for classification is used. Audio deepfakes are detected using CNN models trained on mel spectrograms derived from the ASVspoof 2019 dataset. For video analysis, a ResNet-based frame feature extractor and LSTM model are used to capture temporal inconsistencies, trained on the Celeb-DF dataset. All detection models are integrated into a single user interface using Streamlit, allowing users to input any media type and receive instant detection results. The system achieves high accuracy across all modalities and provides a practical, scalable solution for deepfake identification.
In the old one or we can say previous deepfake detection or Most existing deepfake detection systems focus on a single domain, such as only audio or only video or only image. In real-world scenarios, however, manipulated content can appear in any format. The motivation behind this project is to build a multiple or separate model for respective , that is one for detect deepfake image , one for detect deepfake audio, one for detect deepfake video . This detection system that can detect deepfakes regardless of the input type and deliver results to the user in a simple and interactive way. The primary goals of this system are: 1. To develop an accurate detection model for each media type that help the users to detect the deepfake image , audio ,video so that user can know the reality hidden behind the content and aware the public and save them from this manipulated content traps and make them responsible person .
Key
Words: Deepfake Detection, Multi-Modal Detection, Audio Forgery, Image Manipulation, Video Deepfakes, CNN, LSTM.
2. To integrate all detection pipelines into a single user interface. Previously the user have to use different UI for detect other content , as there is no single UI where user can give image , audio , video at one place to detect the reality hidden behind the content . So we make the single UI where, user can give the input for respective model and can detect the content or input given by users to know whether it is real or deepfake.
1.INTRODUCTION The rapid advancement of AI-generated synthetic media popularly known as deepfakes has given rise to significant concerns around misinformation, identity theft, and manipulation of public discourse. Deepfakes can take the form of realistic synthetic videos, voice clones, or altered images that are often indistinguishable to the human eye or ear.
3. To build a practical tool for general users, where users become the more informative regarding the deepfake content. This prevent the users from the traps of deepfake content . It is also makes the society more responsive regarding the deepfake content as the society people can check the reality of content and knows the real picture of the contents. The all three model are useful and give the accurate result as per the users input and response time is also less.
In this work, we propose a comprehensive machine learningbased system capable of detecting deepfakes in three distinct modalities: image, audio, and video. Our system is backed by well-known datasets (ASVspoof 2019 and Celeb-DF) and architectures (MTCNN, InceptionResNetV1, ResNet, LSTM), and it includes an intuitive UI built using Streamlit that allows users to upload the input which users wants to detect to check whether it is deepfake or not and respective model predict and gives the result back and display on the UI to users.
© 2025, IRJET
|
Impact Factor value: 8.315
4. Aware the society regarding the deepfake content which is created using the newly emerged technologies like AI and ML. This deepfake detection system gives the users more clarity for the deepfake contents and gives them better idea for prevention and their relatives and other people from the menace situation.
|
ISO 9001:2008 Certified Journal
|
Page 607