Update app.py

app.py

@@ -1,6 +1,7 @@
 import cv2
 import numpy as np
 import gradio as gr
+import matplotlib.pyplot as plt
 
 color_ranges = [
     ([100, 150, 0], [140, 255, 255]),  # Blue range
@@ -70,7 +71,7 @@ def apply_filter(filter_type, input_image=None):
     elif filter_type == "Mehmet":
         return apply_multiple_color_filters(frame)
     elif filter_type == "Picasso":
-        return picasso_effect(frame)
+        return split_and_shuffle_image(frame)
 
 
 
@@ -92,32 +93,32 @@ def apply_multiple_color_filters(image):
     filtered_image = cv2.bitwise_and(image, image, mask=combined_mask)
     return filtered_image 
 
-def picasso_effect(image_array):
-    # Convert the RGB array to an image
-    image = cv2.cvtColor(image_array, cv2.COLOR_RGB2BGR)
-
-    # Apply bilateral filter to smooth the image while keeping edges sharp
-    filtered_image = cv2.bilateralFilter(image, d=9, sigmaColor=75, sigmaSpace=75)
-
-    # Convert to grayscale
-    gray_image = cv2.cvtColor(filtered_image, cv2.COLOR_BGR2GRAY)
-
-    # Apply edge detection
-    edges = cv2.Canny(gray_image, threshold1=100, threshold2=200)
-
-    # Dilate the edges to make them more pronounced
-    dilated_edges = cv2.dilate(edges, kernel=np.ones((3, 3), np.uint8), iterations=1)
-
-    # Invert the edges
-    inverted_edges = cv2.bitwise_not(dilated_edges)
-
-    # Convert edges to 3 channels
-    edges_colored = cv2.cvtColor(inverted_edges, cv2.COLOR_GRAY2BGR)
-
-    # Blend the edges with the original image
-    abstract_image = cv2.addWeighted(filtered_image, 0.7, edges_colored, 0.3, 0)
-
-    return abstract_image
+def split_and_shuffle_image(image):
+    # Determine the height and width of each piece
+    h, w, _ = image.shape
+    h_split, w_split = h // 3, w // 3
+    
+    # Split the image into 9 pieces
+    pieces = []
+    for i in range(3):
+        for j in range(3):
+            piece = image[i * h_split:(i + 1) * h_split, j * w_split:(j + 1) * w_split]
+            pieces.append(piece)
+    
+    # Shuffle the pieces
+    np.random.shuffle(pieces)
+    
+    # Reconstruct the shuffled image
+    shuffled_image = np.zeros_like(image)
+    idx = 0
+    for i in range(3):
+        for j in range(3):
+            shuffled_image[i * h_split:(i + 1) * h_split, j * w_split:(j + 1) * w_split] = pieces[idx]
+            idx += 1
+            
+    return shuffled_image
+
+