test1.py

import cv2
import mediapipe as mp
import numpy as np
import time

# The first test - the TUG test
# Returns the total time in Seconds:
def test1():
    # start the timer
    start_time = time.time()

    mp_drawing = mp.solutions.drawing_utils
    mp_drawing_styles = mp.solutions.drawing_styles
    mp_pose = mp.solutions.pose

    # For webcam input:
    cap = cv2.VideoCapture(0)
    with mp_pose.Pose(
        min_detection_confidence=0.5,
        min_tracking_confidence=0.5) as pose:
  
        # initialize variables
        leftStart_y = 0
        rightStart_y = 0
        leftStart_x = 0
        rightStart_x = 0
        rightPrev_x = 0
        leftPrev_x = 0
        frame = 0
        rightx_distance = 0
        leftx_distance = 0

        while cap.isOpened():
            success, image = cap.read()
            if not success:
                print("Ignoring empty camera frame.")
                # If loading a video, use 'break' instead of 'continue'.
                continue

            # To improve performance, optionally mark the image as not writeable to
            # pass by reference.
            image.flags.writeable = False
            image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
            image_height, image_width, _ = image.shape
            results = pose.process(image)

            if 'landmark' in dir(results.pose_world_landmarks):
                right_shoulder = results.pose_landmarks.landmark[mp_pose.PoseLandmark.RIGHT_SHOULDER]
                right_shoulder_pos = np.array([right_shoulder.x * image_width, right_shoulder.y  * image_height, right_shoulder.z])
                left_shoulder = results.pose_landmarks.landmark[mp_pose.PoseLandmark.LEFT_SHOULDER]
                left_shoulder_pos = np.array([left_shoulder.x * image_width, left_shoulder.y  * image_height, left_shoulder.z])
                rightCurr_x = right_shoulder_pos[0]
                rightCurr_y = right_shoulder_pos[1]
                leftCurr_x = left_shoulder_pos[0]
                leftCurr_y = left_shoulder_pos[1]

                # Set start states
                if (frame == 1):
                    rightStart_y = rightCurr_y
                    rightStart_x = rightCurr_x
                    leftStart_y = leftCurr_y
                    leftStart_x = leftCurr_x
                    rightPrev_x = rightStart_x
                    leftPrev_x = leftStart_x

                # Calculate moved distance in x axis of both shoulders
                rightx_distance += abs(rightCurr_x - rightPrev_x)
                leftx_distance += abs(leftCurr_x - leftPrev_x)

                # If the person is sitting and has moved along the x axis for enough distance, stop the timer
                if (abs(rightCurr_y - rightStart_y) < image_height * 0.05) and (abs(leftCurr_y - leftStart_y) < image_height * 0.05) \
                   and (rightx_distance > image_width * 0.4) and (leftx_distance > image_width * 0.4):
                    total_time = time.time() - start_time # stop the timer, return the total time
                    cap.release()
                    return total_time
                
                leftPrev_x = leftCurr_x
                rightPrev_x = rightCurr_x        

            # Draw the pose annotation on the image.
            image.flags.writeable = True
            image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
            mp_drawing.draw_landmarks(
            image,
            results.pose_landmarks,
            mp_pose.POSE_CONNECTIONS,
            landmark_drawing_spec=mp_drawing_styles.get_default_pose_landmarks_style())

            frame += 1

            # Flip the image horizontally for a selfie-view display.
            cv2.putText(image, "time = " + str(int(time.time() - start_time)), (50, 50), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 0, 0), 2, cv2.LINE_AA)
            cv2.imshow('MediaPipe Pose', image)
            if cv2.waitKey(5) & 0xFF == 27:
                break
    
    cap.release()