Frontend Progress

[e-shen2022]

Read comments for in-depth line by line explanation

1. Import Packages

• OpenCV to open camera and capture real-time webcam frames
• Threading & events to coordinate subject and avatar frames

2. Generate path and create dictionaries for emotion text and avatar images

   

3. Defining/Initializing variables

   

• Define variable for frame captured
• Initialize array to hold index of emotion detected
• Switch event to signal when one of functions finish and should start next
• Stop event used when quit button is pressed

4. Show_subject function

• While program is running, open the webcam and capture the frame
• Use haarcascade classifier that detects face in frame using pretrained info ab facial features
• For each detected face, a rectangle is drawn around it on the frame
• Face image is passed through the pre-trained emotion recognition model
• Index of the highest probability in the prediction array is calculated to determine the predicted emotion class
• Corresponding emotion label is retrieved from the emotion_dict dictionary & displayed in window
• Once frame is processed, wait for 0.3 seconds to account for lag

5. Show_avatar function

• While quit button is not pressed, read the frame, display same emotion text as subject

6. Wrapper functions

• Run stop_thread function when quit button is pressed
• Wrapper_quit to stop all threads and destroy root

7. Main commands

• Create labels to position and pack the human subject, emoji, and buttons