Update README.md

README.md

@@ -29,25 +29,16 @@ It seems like using random crops helped the model to generalize better, however,
 
 ## Performance comparison
 
-We did a small comparison with the current available AI image detectors. Note that these models were not specificly trained on anime images.
-
-| Image        | Nahrawy/AIorNot | umm-maybe/AI-image-detector | Organika/sdxl-detector | mmanikanta/VIT_AI_image_detector  | Ours       |
-|--------------|-----------------|-----------------------------|------------------------|-----------------------------------|------------|
-| ai_1.jpg     | ai (100%)       | human (86%)                 | artificial (100%)      | FAKE (89%)                        | ai (100%)  |
-| ai_2.jpg     | ai (99%)        | human (96%)                 | artificial (100%)      | FAKE (89%)                        | ai (100%)  |
-| ai_3.jpg     | ai (77%)        | human (98%)                 | artificial (100%)      | REAL (100%)                       | ai (100%)  |
-| ai_4.jpg     | real (66%)      | human (100%)                | human (100%)           | REAL (100%)                       | real (100%)|
-| ai_5.jpg     | ai (51%)        | human (99%)                 | artificial (55%)       | REAL (99%)                        | real (65%) |
-| ai_6.jpg     | ai (100%)       | human (98%)                 | artificial (100%)      | FAKE (60%)                        | ai (84%)   |
-| real_1.jpg   | ai (99%)        | human (99%)                 | artificial (100%)      | REAL (98%)                        | ai (55%)   |
-| real_2.jpg   | ai (88%)        | human (100%)                | artificial (100%)      | REAL (100%)                       | real (85%) |
-| real_3.jpg   | ai (95%)        | human (96%)                 | artificial (100%)      | REAL (100%)                       | real (97%) |
-| real_4.jpg   | real (90%)      | human (100%)                | artificial (97%)       | REAL (100%)                       | real (94%) |
-| real_5.jpg   | ai (75%)        | human (100%)                | human (57%)            | REAL (100%)                       | real (100%)|
-| real_6.jpg   | ai (89%)        | human (98%)                 | human (100%)           | REAL (100%)                       | real (99%) |
-| **Accuracy:**| 50%             | 50%                         | 58%                    | **75%**                           | **75%**    |
-
-
+We did a small eval test with ~5000 images on the current available AI image detectors. Note that these models were not specificly trained on anime images.
+
+| Model                                    | Accuracy |
+|------------------------------------------|----------|
+| mmanikanta/VIT_AI_image_detector         | 79,65%   |
+| **legekka/AI-Anime-Image-Detector-ViT**      | **94,68%**   |
+| *legekka/AI-Anime-Image-Detector-HD-ViT WIP* | *94,26%*   |
+| Organika/sdxl-detector                   | 43,29%   |
+| umm-maybe/AI-image-detector              | 75,45%   |
+| Nahrawy/AIorNot                          | 64,74%   |
 
 ## Usage