Added the emotion detector classifier

emotion_detection _classification/Emotion_detector_classifier.py

@@ -0,0 +1,83 @@
+# importing the module
+import cv2
+import random
+import numpy as np
+
+import mtcnn
+
+
+from tensorflow.keras.models import model_from_json
+
+detector = mtcnn.MTCNN()
+
+
+
+json_file = open('model.json', 'r')
+loaded_model_json = json_file.read()
+json_file.close()
+
+loaded_model = model_from_json(loaded_model_json)
+
+print(loaded_model.__class__)
+
+loaded_model.load_weights("model.h5")
+print("Loaded model from disk")
+loaded_model.compile(loss='binary_crossentropy', optimizer='rmsprop', metrics=['accuracy'])
+
+emotions = ['anger','fear','happy','neutral','sad']
+
+name = 'examples/smile.jpeg'
+
+img = cv2.imread(name)
+
+
+gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
+
+
+
+height = img.shape[0]
+width = img.shape[1]
+
+
+faces = detector.detect_faces(img)
+
+if (len(faces)>0):
+
+    for f in faces:
+        x,y,w,h = f['box']
+        cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2)
+
+        roi_gray = gray[y:y+h, x:x+w]
+
+        roi_gray = cv2.resize(roi_gray, (224, 224))
+
+        face = cv2.cvtColor(roi_gray,cv2.COLOR_GRAY2RGB)
+
+        # converting image data into proper shape for tensorflow model
+
+        face = np.expand_dims(face, axis=0) 
+
+        emote = loaded_model.predict(face)[0]  ## check face image with model
+
+        emote = np.argmax(emote)
+
+
+        cv2.putText(img,emotions[emote],(x,y-10),cv2.FONT_HERSHEY_SIMPLEX,1,(0,255,0),2,)
+
+cv2.imshow('img',img)    
+
+
+name = name.split('.')[0]
+
+cv2.imwrite('output-'+name+'.png',img)
+
+cv2.waitKey(0)
+cv2.destroyAllWindows()
+
+
+
+
+
+
+
+

emotion_detection _classification/Emotion_detector_classifier_webcam.py

@@ -0,0 +1,88 @@
+# importing the module
+import cv2
+import random
+import numpy as np
+
+import mtcnn
+
+
+from tensorflow.keras.models import model_from_json
+
+
+cap = cv2.VideoCapture(0) 
+
+detector = mtcnn.MTCNN()
+
+
+
+json_file = open('model.json', 'r')
+loaded_model_json = json_file.read()
+json_file.close()
+
+loaded_model = model_from_json(loaded_model_json)
+
+print(loaded_model.__class__)
+
+loaded_model.load_weights("model.h5")
+print("Loaded model from disk")
+loaded_model.compile(loss='binary_crossentropy', optimizer='rmsprop', metrics=['accuracy'])
+
+
+emotions = ['anger','fear','happy','neutral','sad']
+
+while 1:
+
+    ret, img = cap.read()
+
+
+    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
+
+
+
+    height = img.shape[0]
+    width = img.shape[1]
+
+
+    faces = detector.detect_faces(img)
+
+    if (len(faces)>0):
+
+        for f in faces:
+            x,y,w,h = f['box']
+            cv2.rectangle(img,(x,y),(x+w,y+h),(,0,0),2)
+
+            roi_gray = gray[y:y+h, x:x+w]
+
+            roi_gray = cv2.resize(roi_gray, (224, 224))
+
+            face = cv2.cvtColor(roi_gray,cv2.COLOR_GRAY2RGB)
+
+
+            # converting image data into proper shape for tensorflow model
+
+            face = np.expand_dims(face, axis=0) 
+
+            emote = loaded_model.predict(face)[0]  ## check face image with model
+
+            emote = np.argmax(emote)
+
+
+            cv2.putText(img,emotions[emote],(x,y-10),cv2.FONT_HERSHEY_SIMPLEX,1,(0,0,255),2,)
+
+    cv2.imshow('img',img)    
+
+
+    k = cv2.waitKey(30) & 0xff 
+    if k == ord('x'):
+        break
+
+cv2.waitKey(0)
+cv2.destroyAllWindows()
+
+
+
+
+
+
+
+

emotion_detection _classification/README.md

@@ -0,0 +1,8 @@
+Face detection and Emotion classification using MTCNN and EfficientNet B0
+
+Dataset used FER 2013 https://www.kaggle.com/msambare/fer2013
+
+Folders for Surprise and Disgust were removed from both Test and Train folders before training
+
+Document link: https://docs.google.com/document/d/1RxdJks12F8r_bM4kFxoAcud1aGchy5qwwGFv2JLhZXg/edit?usp=sharing
+