FER/detectfaces.py

import cv2
import torch
import os
import time
from PIL import Image

currtime = time.strftime("%H:%M:%S")
print(currtime)


def fer(model_path, device, model):
    # Load the model checkpoint if it exists
    if model_path is not None:
        if os.path.isfile(model_path):
            print("=> loading checkpoint '{}'".format(model_path))
            checkpoint = torch.load(model_path, map_location=device, weights_only=False)
            best_acc = checkpoint["best_acc"]
            best_acc = best_acc.to()
            print(f"best_acc:{best_acc}")
            model.load_state_dict(checkpoint["state_dict"])
            print(
                "=> loaded checkpoint '{}' (epoch {})".format(
                    model_path, checkpoint["epoch"]
                )
            )
        else:
            print(
                "[!] detectfaces.py => no checkpoint found at '{}'".format(model_path)
            )
        # Start webcam capture and prediction
        imagecapture(model)
        return


def imagecapture(model):
    # Initialize webcam capture
    cap = cv2.VideoCapture(0)
    time.sleep(5)  # Wait for 5 seconds to allow the camera to initialize

    # Keep trying to open the webcam until successful
    while not cap.isOpened():
        time.sleep(2)  # Wait for 2 seconds before retrying

    # Flag to control webcam capture
    capturing = True
    while capturing:
        # Import the predict function from the prediction module
        from prediction import predict

        # Read a frame from the webcam
        ret, frame = cap.read()

        # Handle potential error reading the frame
        if not ret:
            print("Error: Could not read frame.")
            break

        # Convert the frame to grayscale
        gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

        # Detect faces using Haar Cascades
        faces = cv2.CascadeClassifier(
            cv2.data.haarcascades + "haarcascade_frontalface_default.xml"
        ).detectMultiScale(gray, scaleFactor=1.3, minNeighbors=5, minSize=(30, 30))

        # If faces are detected, proceed with prediction
        if len(faces) > 0:
            currtimeimg = time.strftime("%H:%M:%S")
            print(f"[!]Face detected at {currtimeimg}")
            # Crop the face region
            face_region = frame[
                faces[0][1] : faces[0][1] + faces[0][3],
                faces[0][0] : faces[0][0] + faces[0][2],
            ]
            # Convert the face region to a PIL image
            face_pil_image = Image.fromarray(
                cv2.cvtColor(face_region, cv2.COLOR_BGR2RGB)
            )
            print("[!]Start Expressions")
            # Record the prediction start time
            starttime = time.strftime("%H:%M:%S")
            print(f"-->Prediction starting at {starttime}")
            # Perform emotion prediction
            predict(model, image_path=face_pil_image)
            # Record the prediction end time
            endtime = time.strftime("%H:%M:%S")
            print(f"-->Done prediction at {endtime}")

            # Stop capturing once prediction is complete
            capturing = False

        # Exit the loop if the 'q' key is pressed
        if cv2.waitKey(1) & 0xFF == ord("q"):
            break

    # Release webcam resources and close OpenCV windows
    cap.release()
    cv2.destroyAllWindows()