face exp

app/Hackathon_setup/exp_model.t7

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:05935a6dbaf7dc4656615384a5f99bf4c665d50195db06725026cc9440589f4d
+size 554893562

app/Hackathon_setup/exp_recognition.py

@@ -65,7 +65,20 @@ def get_expression(img):
     face = detected_face(img)
     if face==0:
         face = Image.fromarray(cv2.cvtColor(img, cv2.COLOR_BGR2GRAY))
-    
+    face = trnscm(face).unsqueeze(0)
     # YOUR CODE HERE, return expression using your model
+    file = open(current_path + "/exp_model.t7")
+    file.seek(0, os.SEEK_END)
+    print("Size of file is :", file.tell(), "bytes")
+    exp_net = facExpRec()
+    model = torch.load(current_path + "/exp_model.t7", map_location="cpu")
+    exp_net.load_state_dict(model["exp_dict"])  ##
 
-    return "YET TO BE CODED"
+    # YOUR CODE HERE, return similarity measure using your model
+    exp_net.eval()
+    face = face.to("cpu")
+    outputs = exp_net(face)
+    _, predicted = torch.max(outputs.data, 1)
+    classes = {0: 'ANGER', 1: 'DISGUST', 2: 'FEAR', 3: 'HAPPINESS', 4: 'NEUTRAL', 5: 'SADNESS', 6: 'SURPRISE'}
+    strexp = classes[predicted.item()]
+    return strexp

app/Hackathon_setup/exp_recognition_model.py

@@ -2,6 +2,7 @@ import torch
 import torchvision
 import torch.nn as nn
 from torchvision import transforms
+import torchvision.models as models
 ## Add more imports if required
 
 ####################################################################################################################
@@ -15,12 +16,25 @@ classes = {0: 'ANGER', 1: 'DISGUST', 2: 'FEAR', 3: 'HAPPINESS', 4: 'NEUTRAL', 5:
 # Example Network
 class facExpRec(torch.nn.Module):
     def __init__(self):
-        pass   # remove 'pass' once you have written your code
-        #YOUR CODE HERE
+
+        super().__init__()
+        self.model = models.vgg16(pretrained=True)
+        for params in self.model.parameters():
+            params.requires_grad = False
+        input_shape = self.model.classifier[6].in_features
+
+        self.model.classifier[6] = nn.Sequential(nn.Linear(input_shape, 1024),nn.ReLU(), nn.Dropout(0.2),
+                                                 nn.Linear(1024,256), nn.ReLU(),nn.Dropout(0.2),
+                                                 nn.Linear(256,7), nn.LogSoftmax(dim=1))
+        self.model.classifier[6].requires_grad = True
+        print("New Layers Added:")
+        for params in self.model.parameters():
+            if params.requires_grad:
+                print(params.shape)
+
         
     def forward(self, x):
-        pass   # remove 'pass' once you have written your code
-        #YOUR CODE HERE
+        return self.model(x)
         
 # Sample Helper function
 def rgb2gray(image):
@@ -28,4 +42,8 @@ def rgb2gray(image):
     
 # Sample Transformation function
 #YOUR CODE HERE for changing the Transformation values.
-trnscm = transforms.Compose([rgb2gray, transforms.Resize((48,48)), transforms.ToTensor()])
+#trnscm = transforms.Compose([rgb2gray, transforms.Resize((48,48)), transforms.ToTensor()])
+trnscm = transforms.Compose([rgb2gray, transforms.Resize((100,100)),
+                                  transforms.ToTensor(),
+                                  transforms.Normalize((0.5,0.5,0.5), (0.5,0.5,0.5))])
+