Abstract:
"The face image manipulation detection system proposed in this research combines deep learning and computer
vision methods to produce accurate and dependable findings. The system begins by extracting different features
from the facial image, such as color, texture, and geometry data. Then, using these attributes, a machine learning
classifier is trained to distinguish between modified and unmanipulated images.
In addition to identifying manipulated photos, the system also carries out localization by determining the specific
areas of the image that have been manipulated. This is accomplished by using a localizing algorithm, which takes
the retrieved data and applies a heatmap display to emphasize the modified areas.
The usefulness of the suggested system was tested through experiments, and the findings show that it is highly
accurate in spotting faked facial photos. In a range of situations, including those involving small modifications
and those involving sophisticated manipulation techniques, the system can successfully recognize manipulated
images.
In conclusion, the technique for detecting and localizing modified face photos that was provided in this paper is
useful. The system is highly suited for a wide range of applications in the disciplines of digital forensics,
biometrics, and security thanks to the combination of modern image processing algorithms and machine learning
models."
