Create app.py

app.py

@@ -0,0 +1,163 @@
+import cv2
+import numpy as np
+import face_recognition
+import os
+import pickle
+from deepface import DeepFace
+import gradio as gr
+from datetime import datetime
+import fast_colorthief
+import webcolors
+from PIL import Image
+
+data1 = []
+try:
+  data = pickle.loads(open("encodings.pickle", "rb").read())
+  data1 = pickle.loads(open("encodings.pickle", "rb").read())
+  data = np.array(data)
+  encodings = [d["encoding"] for d in data]
+  personNames=[d["name"] for d in data]
+except:
+  data=[]
+  data1=[]
+  encodings=[]
+  personNames=[]
+
+# cluster the embeddings
+print("[INFO] clustering...")
+
+print('All Encodings Complete!!!')
+
+
+thres = 0.45
+path = 'images'
+images = []
+classNames= []
+classFile = 'coco.names'
+with open(classFile,'rt') as f:
+  #classNames = f.read().rstrip('n').split('n')
+  classNames = f.readlines()
+
+
+# remove new line characters
+classNames = [x.strip() for x in classNames]
+print(classNames)
+configPath = 'ssd_mobilenet_v3_large_coco_2020_01_14.pbtxt'
+weightsPath = 'frozen_inference_graph.pb'
+net = cv2.dnn_DetectionModel(weightsPath,configPath)
+net.setInputSize(320,320)
+net.setInputScale(1.0/ 127.5)
+net.setInputMean((127.5, 127.5, 127.5))
+net.setInputSwapRB(True)
+
+
+
+def faceEncodings(images):
+    
+    encodeList = []
+    for img in images:
+        img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
+        encode = face_recognition.face_encodings(img)[0]
+        encodeList.append(encode)
+    return encodeList
+
+
+#encodeListKnown = faceEncodings(images)
+
+def main(image): 
+  personsinimage=[]
+  index=len(encodings)
+  gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
+  rgb=cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
+  names=[]
+
+
+  #object
+  try:
+        classIds, confs, bbox = net.detect(image,confThreshold=thres)
+  except  Exception as err:
+      print(err)
+  print(classIds,bbox)
+  try:
+    if len(classIds) != 0:
+        for classId, confidence,box in zip(classIds.flatten(),confs.flatten(),bbox):
+          if names.count(classNames[classId-1]) == 0:
+              names.append(classNames[classId-1])
+  except  Exception as err:
+      print(err)
+  #emotion
+
+  try:
+      face_analysis_2=DeepFace.analyze(image, actions = ['emotion'], enforce_detection=False)
+      names.append(face_analysis_2["dominant_emotion"])
+  except:
+    print("No face")
+    names.append("No Face")
+
+  # #Colour
+
+  colourimage = Image.fromarray(image)
+  colourimage = colourimage.convert('RGBA')
+  colourimage = np.array(colourimage).astype(np.uint8)
+  palette=fast_colorthief.get_palette(colourimage)
+
+
+  for i in range(len(palette)):
+    diff={}
+    for color_hex, color_name in webcolors.CSS3_HEX_TO_NAMES.items():
+      r, g, b = webcolors.hex_to_rgb(color_hex)
+      diff[sum([(r - palette[i][0])**2,
+                (g - palette[i][1])**2,
+                (b - palette[i][2])**2])]= color_name
+    if names.count(diff[min(diff.keys())])==0:
+      names.append(diff[min(diff.keys())])
+
+
+  faces = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
+
+  facesCurrentFrame = face_recognition.face_locations(faces)
+  encodesCurrentFrame = face_recognition.face_encodings(faces, facesCurrentFrame)
+
+  for encodeFace, faceLoc in zip(encodesCurrentFrame, facesCurrentFrame):
+      
+      if len(data1)>0:
+        matches = face_recognition.compare_faces(encodings, encodeFace)
+        faceDis = face_recognition.face_distance(encodings, encodeFace)
+        # print(faceDis)
+        matchIndex = np.argmin(faceDis)
+
+        if matches[matchIndex]:
+            name = personNames[matchIndex].upper()
+            personsinimage.append(name)
+            names.append(name)
+        else:
+          d = [{"name": "unknown"+str(index), "loc": faceLoc, "encoding": encodeFace}]
+          encodings.append(encodeFace)
+          names.append("unknown"+str(index))
+          personNames.append("unknown"+str(index))
+          index+=1
+          data1.extend(d)
+          personsinimage.append("unknown"+str(index))
+      else:
+          d = [{"name": "unknown"+str(index), "loc": faceLoc, "encoding": encodeFace}]
+          encodings.append(encodeFace)
+          names.append("unknown"+str(index))
+          personNames.append("unknown"+str(index))
+          index+=1
+          data1.extend(d)
+
+  # dump the facial encodings data to disk
+  print("[INFO] serializing encodings...")
+  f = open("encodings.pickle", "wb")
+  f.write(pickle.dumps(data1))
+  f.close()
+
+  return ' '.join(names)
+interface = gr.Interface(fn=main, 
+                        inputs=["image"],
+                        outputs=[gr.inputs.Textbox(label='Names of person in image')], 
+                        title='Face Recognition ',
+                        description='This Space:\n1) Recognize face \n2) Detect Emotion \n3) Detect Colors.\n4) Object Detection \n<br>This space is for personal use. For the time being, it only recognizes few persons. To use it, simply duplicate the space and replace the images in the image folder.')
+                        
+
+interface.launch(inline=False)