ViT-Based Copy-Move Forgery Detection Model
Budget: ₹1,000 – ₹3,000 INR
I'm looking to build a model that can detect copy-move forgeries using the multi-scale feature extraction power of Vision Transformers (ViTs). The model should be able to handle complex post-processing, and also be capable of cross-image forgery detection by learning invariant features across different images.
Key Requirements:
- Utilize public dataset for training and testing.
- Ensure robustness against specific post-processing techniques such as blurring, noise addition, and geometric transformations.
- Prioritize accuracy above all else.
Ideal Skills:
- Expertise in Vision Transformers and feature matching mechanisms.
- Previous experience with designing models for forgery detection.
- Proficiency in working with public datasets, CASIA, COVERAGE.
Key Requirements:
- Utilize public dataset for training and testing.
- Ensure robustness against specific post-processing techniques such as blurring, noise addition, and geometric transformations.
- Prioritize accuracy above all else.
Ideal Skills:
- Expertise in Vision Transformers and feature matching mechanisms.
- Previous experience with designing models for forgery detection.
- Proficiency in working with public datasets, CASIA, COVERAGE.