Merge pull request #11 from Em-Vi/add-face-blur-tool

Add: [AI-ML] faceBlurTool

Author: Tejas-Santosh-Nalawade

Domains/AI-ML/MiniProjects/faceBlurTool/README.md

@@ -0,0 +1,41 @@
+import cv2
+from cv2 import data
+
+face_cascade = cv2.CascadeClassifier(data.haarcascades + "haarcascade_frontalface_default.xml")
+
+image_path = input("Enter image path: ")
+image = cv2.imread(image_path)
+
+if image is None:
+    print("Image not found")
+    exit()
+
+# grayscale for better detection
+gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
+
+# lets find faces
+# cascade detector runs on th eimage and returns a rectangle coordinate x,y,w,h
+# scalefactor repatedly scales image by value to detet faces in differnet sizes
+# minneighbour decides how many rectabgles should be overlapped to trigger face detection
+# minsize ignores detected objects smaller tha 30x30 pixels
+faces = face_cascade.detectMultiScale(gray,scaleFactor=1.1,minNeighbors=5,minSize=(30,30))
+
+#looping over detected face coords
+for (x,y,w,h) in faces:
+    # extacts the subarray containing detected face
+    face = image[y:y+h,x:x+w]
+    # numpy doesnt use a copy but views into image
+    # lets apply guassian blur
+    blurred_face = cv2.GaussianBlur(face,(99,99),30)
+    # (99, 99) is the kernel size. stronger blr
+    # 30 is sigmaX (standard deviation in the X direction). used to make rexziabke with face sizes
+    # we can write blurred_face in original image
+    image[y:y+h, x:x+w] = blurred_face
+    
+output_path = "blurred_" + image_path.split("/")[-1]
+cv2.imwrite(output_path,image)
+# show image...comment out if not needed
+cv2.imshow("Blurred Image", image)
+cv2.waitKey(0)
+cv2.destroyAllWindows()
+