International Journal of Computer Applications
Foundation of Computer Science (FCS), NY, USA
|
Volume 186 - Issue 57 |
Published: December 2024 |
Authors: Nandini S., Chethesh B.L. |
![]() |
Nandini S., Chethesh B.L. . Harnessing Deep Learning for Reliable Detection of DeepFake Images. International Journal of Computer Applications. 186, 57 (December 2024), 7-12. DOI=10.5120/ijca2024924298
@article{ 10.5120/ijca2024924298, author = { Nandini S.,Chethesh B.L. }, title = { Harnessing Deep Learning for Reliable Detection of DeepFake Images }, journal = { International Journal of Computer Applications }, year = { 2024 }, volume = { 186 }, number = { 57 }, pages = { 7-12 }, doi = { 10.5120/ijca2024924298 }, publisher = { Foundation of Computer Science (FCS), NY, USA } }
%0 Journal Article %D 2024 %A Nandini S. %A Chethesh B.L. %T Harnessing Deep Learning for Reliable Detection of DeepFake Images%T %J International Journal of Computer Applications %V 186 %N 57 %P 7-12 %R 10.5120/ijca2024924298 %I Foundation of Computer Science (FCS), NY, USA
The Deepfake software has developed as a potent tool for creating extremely realistic but deceptive graphics, presenting serious safety and security issues. As fake information algorithms advance, differentiating both actual and modified images gets more difficult This study addresses this concern by using a powerful analysis algorithm that uses neural networks to distinguish between real and manipulated images Specifically, three convolutional neural networks— XceptionNet, InceptionV3, and EfficientNetB0— is used for this task. The simulation is performed using a set of data with the changed facial characteristics, including eyes, mouth, mid-face, and nose First-order techniques such as shrinkage and standardization are used to provide a model the performance is improved. This technology analyzes images as "real" or "fake" and detects changing facial feature areas. The model performance is measured using various metrics which includes accuracy, precision, recall, and confusion matrices, enabling appropriate and efficient depth feature detection