Emotion Detector using Google’s Teachable ML

Realtime Emotion Detector using Python (Google’s Teachable Machine Learning)

🛠️ Description

This project is about developing a system that can detect the emotions of a person in realtime from the video on the basis of a pre-trained keras model. This model was trained using Google’s Teachable Machine Learning.

The project can detect the following emotions:- Angry, Happy, Sad, Smile, Surprise

⚙️ Languages or Frameworks Used

• Python, Mediapipe, Keras
• Teachable Machine Learning (For model training)

🌟 How to run

• Install all the requirements

  Run pip install -r requirements.txt to install all the requirements.

• Run the project

  To the run the project, go to the terminal and run python main.py. This will popup two windows, one for capturing the video input and the other for displyaing the emotion output.

Note: The Model (.h5 file) has been trained using the Teachable Machine Learning which is an esay to use ML Training Platform by Google. Do checkout that platform.

📺 Demo

Do checkout the Below Video for Demo of the Project.

Youtube Link

🤖 Author

Github - MBSA-INFINITY LinkedIn - MBSAIADITYA Portfolio - MBSA Instagram - MBSAIADITYA

Source Code: main.py

import cv2
import numpy as np
import mediapipe as mp
from keras.models import  load_model
from keras.preprocessing import image
# from tensorflow.keras.utils import img_to_array
from PIL import Image, ImageOps

mpFaceDetection = mp.solutions.face_detection
mpDraw = mp.solutions.drawing_utils
faceDetection = mpFaceDetection.FaceDetection()

model = load_model('./Teachable ML Data/keras_model.h5')

cap = cv2.VideoCapture(1)

results_detect = {0:"😁",1:"😠",2:"☹️",3:"😊",4:"😲"}
results_detect_str = {0:"happy",1:"angry",2:"sad",3:"smile",4:"surprise"}

# pTime = 0
while cap.isOpened():
    _,img = cap.read()
    imgRGB = cv2.cvtColor(img,cv2.COLOR_BGR2RGB)
    results = faceDetection.process(imgRGB)
    if results.detections:
        ih,iw,ic = img.shape
        for id,detection in enumerate(results.detections):
            bBoxC = detection.location_data.relative_bounding_box
            bBox = int(bBoxC.xmin * iw),int(bBoxC.ymin * ih),int(bBoxC.width * iw),int(bBoxC.height * ih)
            # cv2.rectangle(img,bBox,(255,0,255),2)
            roi_gray = img[bBox[1]:bBox[1] + bBox[2], bBox[0]:bBox[0] + bBox[3]]
            roi_gray = cv2.resize(roi_gray, (224, 224))
            cv2.imwrite("image.jpg",roi_gray)


            data = np.ndarray(shape=(1, 224, 224, 3), dtype=np.float32)
            image = Image.open('image.jpg')
            size = (224, 224)
            image = ImageOps.fit(image, size, Image.ANTIALIAS)
            image_array = np.asarray(image)
            normalized_image_array = (image_array.astype(np.float32) / 127.0) - 1
            data[0] = normalized_image_array
            prediction = model.predict(data)
            res = np.argmax(prediction)

            # predictions = np.argmax(model.predict(np.array([roi_gray])))
            # cv2.putText(img, results_detect[res], (150,50), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 0, 255), 2)

            temp_emotion = cv2.imread(f"./emotions/{results_detect_str[res]}.jfif")
            cv2.imshow("emotion", temp_emotion)
            print(results_detect[res])

    cv2.imshow("Image",img)

    key = cv2.waitKey(1)
    if key == ord('q'):
        cv2.destroyAllWindows()
        break