update links and README

README.md

@@ -87,7 +87,7 @@ By default, datasets are loaded over the network based on `backend/config.py`. P
 
 There is commented out code that sets the `ABS_DATASET_DOMAIN` to the `backend/data` directory. You can see the data formats of the csv / json files required to render the leaderboard as well as examples there.
 
-In the `data` directory, by default this matches the path structure for loading over the network. Each dataset should be placed under `data/omnisealbench` as a directory, e.g. `data/omnisealbench/sav_val_full_v2` and in the directory have files:
+In the `data` directory, by default this matches the path structure for loading over the network. Each dataset should be placed under `data/omnisealbench` as a directory, e.g. `data/omnisealbench/sav_val_full_v3` and in the directory have files:
 
 - `{type}_benchmark.csv` for leaderboard tables
 - `{type}_attacks_variations.csv` for leaderboard chart

backend/descriptions.py

@@ -210,7 +210,7 @@ MODEL_DESCRIPTIONS = {
   "pixelseal": {
     "full_name": "PixelSeal",
     "description": "Adversarial-only training for invisible image and video watermarking.",
-    "paper_link": "https://arxiv.org/abs/2412.09492",
+    "paper_link": "https://arxiv.org/abs/2512.16874",
     "github_link": "https://github.com/facebookresearch/videoseal",
     "message_size": "256 bits",
     "alias": "PixelSeal"