detection + robustness info

README.md

@@ -24,4 +24,66 @@ vae = AutoencoderKL.from_pretrained("imatag/stable-signature-bzh-sdxl-vae-medium
 pipe = StableDiffusionXLPipeline.from_pretrained(model, vae=vae)
 ```
 
-For more information, please have a look at [the official demo](https://huggingface.co/spaces/imatag/stable-signature-bzh)
+For more information, please have a look at [the official demo](https://huggingface.co/spaces/imatag/stable-signature-bzh)
+
+## Decoder Finetuning
+
+This model was finetuned from the original StableDiffusion-XL autoencoder with the additional objective of including a medium invisible watermark, following the procedure of [StableSignature](https://ai.meta.com/blog/stable-signature-watermarking-generative-ai/). This watermark is detectable by [IMATAG](https://www.imatag.com/)'s demo BZH decoder, available via API.
+
+## Detecting the watermark
+
+For security reasons we cannot release the weights of the detector. Instead, you may check for the presence of a watermark with the [detect_api.py](https://huggingface.co/spaces/imatag/stable-signature-bzh/resolve/main/detect_api.py) script like this:
+```shell
+python detect_api.py test.png
+```
+
+### Visual
+_Visualization of watermark impact on 512x512 images from the COCO2017 validation dataset prompts with SDXL-turbo._ 
+
+<p align="center">
+  <br>
+  <b>
+512x512: original (left), watermarked (middle), difference (right)</b>
+</p>
+
+<p align="center">
+<img src=https://huggingface.co/spaces/imatag/stable-signature-bzh/resolve/main/eval/medium/0001.png />
+</p>
+
+<p align="center">
+<img src=https://huggingface.co/spaces/imatag/stable-signature-bzh/resolve/main/eval/medium/0003.png />
+</p>
+
+<p align="center">
+<img src=https://huggingface.co/spaces/imatag/stable-signature-bzh/resolve/main/eval/medium/0004.png />
+</p>
+
+<p align="center">
+<img src=https://huggingface.co/spaces/imatag/stable-signature-bzh/resolve/main/eval/medium/0005.png />
+</p>
+
+<p align="center">
+<img src=https://huggingface.co/spaces/imatag/stable-signature-bzh/resolve/main/eval/medium/0006.png />
+</p>
+
+### Robustness
+
+The watermark is robust to most editorial changes to an image, such as cropping, changing brightness or contrast, resizing or JPEG compression:
+
+<p align="center">
+  <br>
+  <b>
+from left to right: original, watermarked, difference, 'crop 50% + brigthen 50% + jpeg 80%' attack, 'downscale 2x + crop 50% + brigthen 50% + jpeg 50%' attack</b>
+</p>
+
+<p align="center">
+<img src=https://huggingface.co/spaces/imatag/stable-signature-bzh/resolve/main/eval/medium/0000.png />
+</p>
+
+The Receiver Operating Characteristic curve shows its performance under no modification and when modified by the two scenarios illustrated above:
+<p align="center">
+<img src=https://huggingface.co/spaces/imatag/stable-signature-bzh/resolve/main/eval/medium/roc.png />
+</p>
+
+More details available in our [announcement](https://www.imatag.com/blog/unlocking-the-future-of-content-authentication-imatags-breakthrough-in-ai-generated-image-watermarking) and our lab's [blog post](https://imatag-lab.medium.com/stable-signature-meets-bzh-53ad0ba13691).
+For watermarked models with a different key, support for payload, other perceptual compromises, robustness to other attacks, or faster detection, please [contact IMATAG](https://pages.imatag.com/contact-us-imatag).