FAI.CE Team Submission

.gitignore

@@ -0,0 +1 @@
+venv

README.md

@@ -1,7 +1,21 @@
-# FER May Hakathon
+# Team FAI.CE: FER May Hakathon
 
 Facial Emotion Detection Hackathon Project, Create a model and test it uing 5 to 10 sec videos to detect emotions 
 
+Team Members:
+Mohamed Ratiq
+Bhavika Kaliya
+Alora Tabuco
+
+## Run Locally
+streamlit run app.py --server.enableXsrfProtection false
+(This is only to allow file uploads when running locally)
+
+## View Deployment Link
+https://fer-may-hackathon-faice.streamlit.app/ 
+
+Please note that due to limitations with streamlit cloud - the performance is relatively slow on deployment.
+
 # Facial Emotion Recognition
 
 <div id="top"></div>

app.py

@@ -0,0 +1,90 @@
+import streamlit as st
+import cv2
+import tempfile
+import os
+import numpy as np
+import tensorflow as tf
+from tensorflow.keras.models import load_model
+
+# This can be changed between cnnModel.h5, emotion_recognition_model.h5, and mobilenet.h5
+model_path = 'mobilenet.h5'
+emotion_model = load_model(model_path)
+
+# Emotion labels
+emotion_labels = ['Angry', 'Disgust', 'Fear', 'Happy', 'Sad', 'Surprise', 'Neutral']
+
+def main():
+    # Title of the app
+    st.title("Video Input App with Face Detection")
+
+    with st.expander("Demo Video"):
+        st.video('assets/demo.webm')
+
+    # File uploader for video input
+    video_file = st.file_uploader("Upload a video file", type=["mp4", "mov", "avi"])
+
+    if video_file is not None:
+        # Create a temporary file to save the uploaded video
+        tfile = tempfile.NamedTemporaryFile(delete=False)
+        tfile.write(video_file.read())
+
+        # Play the video and perform face detection
+        st.write("Processing video for face detection...")
+        process_video(tfile.name)
+
+        # Clean up: remove the temporary file
+        tfile.close()
+        os.unlink(tfile.name)
+
+def process_video(video_path):
+    # Load OpenCV's pre-trained Haar Cascade face detector
+    face_cascade = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_frontalface_default.xml')
+
+    video_capture = cv2.VideoCapture(video_path)
+
+    stframe = st.empty()
+
+    while video_capture.isOpened():
+        ret, frame = video_capture.read()
+        if not ret:
+            break
+
+        # Convert the frame to grayscale
+        gray_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
+
+        # Detect faces
+        faces = face_cascade.detectMultiScale(gray_frame, scaleFactor=1.1, minNeighbors=5, minSize=(30, 30))
+
+        # Draw rectangles around the faces
+        for (x, y, w, h) in faces:
+            cv2.rectangle(frame, (x, y), (x + w, y + h), (0, 255, 0), 2)
+
+            # Extract the face ROI
+            face = gray_frame[y:y+h, x:x+w]
+            # Resize the face to 48x48 pixels
+            face_resized = cv2.resize(face, (224, 224))
+            # Normalize the pixel values
+            face_normalized = face_resized / 255.0
+            # Expand dimensions to match model input shape
+            face_rgb = np.stack((face_normalized,) * 3, axis=-1)
+            face_input = np.expand_dims(face_rgb, axis=0)
+
+            # Predict the emotion
+            emotion_prediction = emotion_model.predict(face_input)
+            emotion_label = emotion_labels[np.argmax(emotion_prediction)]
+
+            # Draw a rectangle around the face and put the emotion label
+            cv2.rectangle(frame, (x, y), (x + w, y + h), (0, 255, 0), 2)
+            cv2.putText(frame, emotion_label, (x, y - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.9, (0, 255, 0), 2)
+
+
+        # Convert the frame back to RGB
+        rgb_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
+
+        # Display the frame
+        stframe.image(rgb_frame, channels="RGB")
+
+    video_capture.release()
+
+if __name__ == "__main__":
+    main()