PS|WIP-facial texture

app.py

@@ -1,10 +1,11 @@
 import gradio as gr
+from cv_utils.facial_texture import compute_face_simplicity
 
 def identity_function(input_image):
     return input_image
 
 iface = gr.Interface(
-    fn=identity_function,
+    fn=compute_face_simplicity,
     inputs=gr.inputs.Image(type="pil"),
     outputs=gr.outputs.Image(type="pil")
 )

cv_utils/facial_texture.py

@@ -0,0 +1,77 @@
+import cv2
+import numpy as np
+from skimage.feature import local_binary_pattern
+import matplotlib.pyplot as plt
+import dlib
+import imutils
+import os
+from PIL import Image
+
+
+def compute_face_simplicity(image):
+    
+    ######## create if or depending on input - filepath or PIL file
+    # Load the image from a filepath
+    # image = cv2.imread(image_path)
+    
+    # Convert RGB to BGR format (OpenCV uses BGR while PIL uses RGB)
+    image_np = np.array(image)
+    image = cv2.cvtColor(image_np, cv2.COLOR_RGB2BGR)
+
+    # Resize the image
+    image = imutils.resize(image, width=800)
+
+    # Convert to grayscale
+    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
+    
+    detector = dlib.get_frontal_face_detector()
+    predictor = dlib.shape_predictor("models/face_alignment/shape_predictor_68_face_landmarks.dat")
+    
+    # Detect the face in the image
+    faces = detector(gray, 1)
+    if len(faces) == 0:
+        return "No face detected."
+
+    x, y, w, h = (faces[0].left(), faces[0].top(), faces[0].width(), faces[0].height())
+    face_img = gray[y:y+h, x:x+w]
+    
+    
+    # Parameters for LBP
+    radius = 1
+    n_points = 8 * radius
+    
+    # Apply LBP to the face region
+    lbp = local_binary_pattern(face_img, n_points, radius, method="uniform")
+    
+    # Compute the histogram of the LBP
+    hist, _ = np.histogram(lbp.ravel(), bins=np.arange(0, n_points + 3), range=(0, n_points + 2))
+    
+    # Measure the variance of the histogram
+    variance = np.var(hist)
+    std = np.sqrt(variance)
+    print(std)
+    
+    # A hypothetical threshold - needs calibration
+    threshold = 10000
+    
+    if std < threshold:
+        simplicity = "Simple"
+    else:
+        simplicity = "Complex"
+    
+    # Visualizing the LBP pattern on the detected face
+    # plt.imshow(lbp)
+    lbp = (lbp * 255).astype(np.uint8)
+    lbp = Image.fromarray(lbp)
+    
+    return lbp #, simplicity
+
+
+if __name__ == "__main__":
+    print(os.getcwd())
+    detector = dlib.get_frontal_face_detector()
+    predictor = dlib.shape_predictor("models/face_alignment/shape_predictor_68_face_landmarks.dat")
+    print(predictor)
+    
+    image_path = 'data/images_symmetry/gigi_hadid.webp'
+    print(compute_face_simplicity(image_path))