🏗️ build(utils): update image preprocess and functionality expansion

- move ELA functionality to separate utils for better modulation and maintenance【config】

🏗️ build(utils): update image preprocess and functionality expansion

- update ELA for better performance【Install libraries Pytorch, PIL and opencv】

🔧 chore(doc): update py files

- remove ON.NX dependencies【Remove onnx pytorch dependencies】

🔧 chore(doc): update py files

- update ignore files setting to educate CI autoignore build files

.gitignore

@@ -1,3 +1,5 @@
 .env
 *goat.py
-.vscode
+.vscode
+*onnx.py
+./models/*

app.py

@@ -10,9 +10,10 @@ from PIL import Image
 import numpy as np
 import io
 import logging
-from utils.utils import softmax, augment_image, convert_pil_to_bytes, ELA
+from utils.utils import softmax, augment_image, convert_pil_to_bytes
 from utils.gradient import gradient_processing
 from utils.minmax import preprocess as minmax_preprocess
+from utils.ela import getELA as ELA
 
 
 # Configure logging
@@ -212,7 +213,7 @@ def generate_results_html(results):
                     <svg xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 24 24" stroke-width="3" stroke="currentColor" class="w-4 h-4 mr-2 -ml-3 group-hover:animate group-hover:animate-pulse">
                         {'<path stroke-linecap="round" stroke-linejoin="round" d="M9 12.75 11.25 15 15 9.75M21 12a9 9 0 1 1-18 0 9 9 0 0 1 18 0Z" />' if label == 'REAL' else '<path stroke-linecap="round" stroke-linejoin="round" d="m9.75 9.75 4.5 4.5m0-4.5-4.5 4.5M21 12a9 9 0 1 1-18 0 9 9 0 0 1 18 0Z" />'}
                     </svg>
-                    <p class="whitespace-nowrap text-lg leading-normal font-bold text-center self-center align-middle py-px">{label}</p>
+                    <p class="text-base whitespace-nowrap lg:text-lg leading-normal font-bold text-center self-center align-middle py-px">{label}</p>
                 </span>
             </div>
             <div>

utils/ela.py

@@ -0,0 +1,21 @@
+import numpy as np
+import io
+from PIL import Image, ImageFilter, ImageChops
+from torchvision import transforms
+
+def genELA(img_pil, scale=77, alpha=0.66):
+    # Error Level Analysis for basic image forensics
+    original = img_pil.copy()          # open up the input image
+    temp_path = 'temp.jpg'             # temporary image name to save the ELA to
+    original.save(temp_path, quality=95)     # re-save the image with a quality of 95%
+    temporary = Image.open(temp_path)        # open up the re-saved image
+
+    diff = ImageChops.difference(original, temporary)   # load in the images to look at pixel by pixel differences
+    d = diff.load()                     # load the image into a variable
+    WIDTH, HEIGHT = diff.size           # set the size into a tuple
+    for x in range(WIDTH):                                  # row by row
+        for y in range(HEIGHT):                             # column by column
+            d[x, y] = tuple(k * scale for k in d[x, y])     # set the pixels to their x,y & color based on error
+
+    new_img = ImageChops.blend(temporary, diff, alpha)      # blend the original w/ the ELA @ a set alpha/transparency
+    return new_img

utils/minmax.py

@@ -27,7 +27,7 @@ def blk_filter(img, radius):
             )
     return cv.normalize(result, None, 0, 127, cv.NORM_MINMAX, cv.CV_8UC1)
 
-def preprocess(image, channel=4, radius=1):
+def preprocess(image, channel=4, radius=2):
     if not isinstance(image, np.ndarray):
         image = np.array(image)  # Ensure image is a NumPy array
     if channel == 0:

utils/onnx.py

@@ -1,34 +0,0 @@
-import torch
-from transformers import AutoImageProcessor, Swinv2ForImageClassification
-import onnx
-import onnxruntime as ort
-
-# Load the model and processor
-image_processor = AutoImageProcessor.from_pretrained("haywoodsloan/ai-image-detector-deploy")
-model = Swinv2ForImageClassification.from_pretrained("haywoodsloan/ai-image-detector-deploy")
-
-# Set the model to evaluation mode
-model.eval()
-
-# Create dummy input for tracing
-dummy_input = torch.randn(1, 3, 256, 256)  # Batch size of 1, 3 color channels, 256x256 image
-
-# Export the model to ONNX
-onnx_model_path = "model.onnx"
-# torch.onnx.export(
-#     model,
-#     dummy_input,
-#     onnx_model_path,
-#     input_names=["pixel_values"],
-#     output_names=["logits"],
-#     opset_version=11,
-#     dynamic_axes={
-#         "pixel_values": {0: "batch_size"},
-#         "logits": {0: "batch_size"}
-#     }
-# )
-
-# Verify the ONNX model
-# onnx_model = onnx.load(onnx_model_path)
-# onnx.checker.check_model(onnx_model)
-print("The ONNX model is valid.")

utils/utils.py

@@ -24,19 +24,4 @@ def convert_pil_to_bytes(image, format='JPEG'):
     img_byte_arr = img_byte_arr.getvalue()
     return img_byte_arr
 
-def ELA(img_pil, scale=77, alpha=0.66):
-    # Error Level Analysis for basic image forensics
-    original = img_pil.copy()          # open up the input image
-    temp_path = 'temp.jpg'             # temporary image name to save the ELA to
-    original.save(temp_path, quality=95)     # re-save the image with a quality of 95%
-    temporary = Image.open(temp_path)        # open up the re-saved image
 
-    diff = ImageChops.difference(original, temporary)   # load in the images to look at pixel by pixel differences
-    d = diff.load()                     # load the image into a variable
-    WIDTH, HEIGHT = diff.size           # set the size into a tuple
-    for x in range(WIDTH):                                  # row by row
-        for y in range(HEIGHT):                             # column by column
-            d[x, y] = tuple(k * scale for k in d[x, y])     # set the pixels to their x,y & color based on error
-
-    new_img = ImageChops.blend(temporary, diff, alpha)      # blend the original w/ the ELA @ a set alpha/transparency
-    return new_img