add 3 binarization strategies

app.py

@@ -1,7 +1,104 @@
 import gradio as gr
+import cv2
+import numpy as np
+import torch
+import spaces
+from PIL import Image
 
-def greet(name):
-    return "Hello " + name + "!!"
+# -------------------------------------------------------------
+# 1. SAUVOLA BINARIZATION (Traditional / CPU)
+# -------------------------------------------------------------
+def run_sauvola(image_np, window_size=15, k=0.2, r=128):
+    """
+    Standard Sauvola local thresholding using OpenCV.
+    Formula: T = m * (1 + k * (s / R - 1))
+    """
+    gray = cv2.cvtColor(image_np, cv2.COLOR_RGB2GRAY)
+    window_size = int(window_size) | 1 # Ensure odd window size
+    
+    mean = cv2.blur(gray, (window_size, window_size))
+    mean_sq = cv2.blur(gray**2, (window_size, window_size))
+    std = np.sqrt(mean_sq - mean**2)
+    
+    threshold = mean * (1.0 + k * (std / r - 1.0))
+    binary = np.where(gray > threshold, 255, 0).astype(np.uint8)
+    return Image.fromarray(binary)
 
-demo = gr.Interface(fn=greet, inputs="text", outputs="text")
-demo.launch()
+# -------------------------------------------------------------
+# 2. TZEFA-BINARIZATION (HF Zero GPU)
+# -------------------------------------------------------------
+@spaces.GPU
+def run_tzefa(image_pil):
+    # Load WARAJA/Tzefa-Binarization (uses sbb_binarization / ResNet+Transformer)
+    # Ensure you load the pipeline/model *inside* or cached globally
+    from transformers import pipeline
+    # Example wrapper depending on how Tzefa hosts their pipeline:
+    # pipe = pipeline("image-to-image", model="WARAJA/Tzefa-Binarization")
+    # return pipe(image_pil)
+    pass
+
+# -------------------------------------------------------------
+# 3. TWO-STAGE GAN (opensuh/DocumentBinarization)
+# -------------------------------------------------------------
+@spaces.GPU
+def run_two_stage_gan(image_pil):
+    # 1. Preprocess image to fit the 512x512 patches or required shapes
+    # 2. Pass through Stage 1 (Enhancement Network)
+    # 3. Pass through Stage 2 (Global/Local Binarization Network)
+    pass
+
+# -------------------------------------------------------------
+# 4. DOCRES GENERALIST MODEL (HF Zero GPU)
+# -------------------------------------------------------------
+@spaces.GPU
+def run_docres(image_pil):
+    # DocRes acts as an intelligent outer shell handling unified tasks.
+    # We pass it the image along with its DTSPrompt (Dynamic Task-Specific Prompt) 
+    # tailored specifically for the 'Binarization' task.
+    pass
+
+def process_image(input_img, algo_choice, sauvola_w, sauvola_k):
+    # Convert PIL to Numpy for opencv processing if needed
+    img_np = np.array(input_img)
+    
+    if algo_choice == "Sauvola (Traditional)":
+        return run_sauvola(img_np, sauvola_w, sauvola_k)
+    elif algo_choice == "Tzefa-Binarization":
+        return run_tzefa(input_img)
+    elif algo_choice == "Two-Stage GAN (opensuh)":
+        return run_two_stage_gan(input_img)
+    elif algo_choice == "DocRes (Generalist Transformer)":
+        return run_docres(input_img)
+
+# Building the Interface
+with gr.Blocks(theme=gr.themes.Soft()) as demo:
+    gr.Markdown("# 📄 Document Image Binarization Benchmarking Suite")
+    gr.Markdown("Compare historical document cleaning, GAN-based restoration, and local adaptive thresholding.")
+    
+    with gr.Row():
+        with gr.Column(scale=1):
+            input_image = gr.Image(type="pil", label="Upload Degraded Document")
+            
+            algo = gr.Dropdown(
+                choices=["Sauvola (Traditional)", "Tzefa-Binarization", "Two-Stage GAN (opensuh)", "DocRes (Generalist Transformer)"],
+                value="Sauvola (Traditional)",
+                label="Select Binarization Engine"
+            )
+            
+            # Interactive container for Sauvola params (hides/shows dynamically)
+            with gr.Group() as sauvola_params:
+                s_w = gr.Slider(minimum=3, maximum=99, step=2, value=15, label="Sauvola Window Size")
+                s_k = gr.Slider(minimum=0.0, maximum=1.0, step=0.05, value=0.2, label="Sauvola K Parameter")
+                
+            submit_btn = gr.Button("Binarize Document", variant="primary")
+            
+        with gr.Column(scale=1):
+            output_image = gr.Image(type="pil", label="Binarized / Enhanced Output")
+            
+    submit_btn.click(
+        fn=process_image, 
+        inputs=[input_image, algo, s_w, s_k], 
+        outputs=output_image
+    )
+
+demo.launch()

requirements.txt

@@ -0,0 +1,16 @@
+# Basic interface and image processing
+gradio==4.44.0  
+opencv-python-headless
+numpy
+pillow
+
+# Deep Learning core
+torch
+torchvision
+torchaudio --index-url https://download.pytorch.org/whl/cu121
+transformers
+segmentation-models-pytorch
+pytesseract
+
+# For Hugging Face Zero instances
+spaces