| --- |
| title: XAI - DeepfakeDetection & Forensics Suite |
| emoji: 👁 |
| colorFrom: red |
| colorTo: red |
| sdk: gradio |
| sdk_version: 6.22.0 |
| python_version: '3.12' |
| app_file: app.py |
| pinned: false |
| license: mit |
| models: |
| - buildborderless/CommunityForensics-DeepfakeDet-ViT |
| datasets: |
| - buildborderless/deepfake-explainability |
| short_description: Most comprehensive suite of tests for research and datasets. |
| --- |
| |
| # DeepfakeDet-ViT XAI & Digital Forensics Suite |
|
|
| Interactive explainability dashboard and digital forensics suite for the **ViT-Small Deepfake Classifier** (`buildborderless/CommunityForensics-DeepfakeDet-ViT`). |
|
|
| ## Features |
| - **7 ViT Attribution Methods**: Attention Rollout, Chefer Gradient-Rollout, GradCAM, Integrated Gradients, DeepLIFT, Occlusion Sensitivity, Per-Head Attention. |
| - **5 Classic Digital Forensics Tools**: Error Level Analysis (ELA), Spatial Edge Gradient, Bit Plane Extractor, MinMax Deviation, Wavelet Noise Estimation. |
| - **Forensic Report Generator**: Export multi-panel high-DPI infographic report PNGs. |
| - **Dataset Integration**: Log predictions, user ground truth feedback (Real / Fake / FP / FN), and 7-method composite XAI heatmaps to `buildborderless/deepfake-explainability`. |
