The innovation for preventing fall in elder by using the medical techniques call “Fall risk check list” and "Time up and go test" integrate with the computer vision technique to calculate the risk of fall.

Reference:

AI Pose Estimation with Python and MediaPipe | Plus AI Gym Tracker Project - YouTube

Import library and setup

!pip install mediapipe opencv-python
!pip install gspread
!pip install playsound

import cv2
import mediapipe as mp
import numpy as np
from datetime import datetime
import pandas as pd
import IPython
from IPython.display import Audio
import time
from pathlib import Path
from playsound import playsound

audio = Path().cwd() / "test_2.wav" #your file name

# Times
now = datetime.now().time()
now = (str(now).replace(".", "_").replace(":", "_"))

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

Create a function to calculate the angle

def calculate_angle(a,b,c):
    a = np.array(a) # First
    b = np.array(b) # Mid
    c = np.array(c) # End
    
    radians = np.arctan2(c[1]-b[1], c[0]-b[0]) - np.arctan2(a[1]-b[1], a[0]-b[0])
    angle = np.abs(radians*180.0/np.pi)
    
    if angle >180.0:
        angle = 360-angle
        
    return angle

# Using NumPy
dtypes = np.dtype(
    [
        ("No_Name", str),
        ("Total", str),
        ("Sit Total", str),
        ("Time", str),
        ("Name", str)
    ]
)
df = pd.DataFrame(np.empty(0, dtype=dtypes))#cap = cv2.VideoCapture('tug (2) (1).mp4')
run_no = str(input())

playsound(audio)
time.sleep(1)

cap = cv2.VideoCapture(0)

# Curl counter variables
counter = 0 
stage = None
i=0
start_time = time.time()
time_spare = 0 
test = True
prev_frame_time = 0
new_frame_time = 0
sit_count = []

fourcc = cv2.VideoWriter_fourcc(*'MP4V')
out = cv2.VideoWriter(f'output_{str(now)}.mp4', fourcc, 20.0, (640,  480))

## Setup mediapipe instance
with mp_pose.Pose(min_detection_confidence=0.5, min_tracking_confidence=0.5) as pose:
    while cap.isOpened():

        ret, frame = cap.read()
        

        # Recolor image to RGB
        image = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
        image.flags.writeable = False
      
        # Make detection
        results = pose.process(image)
    
        # Recolor back to BGR
        image.flags.writeable = True
        image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
        
        image_height, image_width, _ =  image.shape
        image = cv2.resize(image, (int(image_width * (480 / image_height)), 480))
        
        # Extract landmarks
        try:
            landmarks = results.pose_landmarks.landmark
            
            # Get coordinates
            shoulder = [landmarks[mp_pose.PoseLandmark.LEFT_HIP.value].x,landmarks[mp_pose.PoseLandmark.LEFT_HIP.value].y]
            elbow = [landmarks[mp_pose.PoseLandmark.LEFT_KNEE.value].x,landmarks[mp_pose.PoseLandmark.LEFT_KNEE.value].y]
            wrist = [landmarks[mp_pose.PoseLandmark.LEFT_WRIST.value].x,landmarks[mp_pose.PoseLandmark.LEFT_WRIST.value].y]
            
            # Calculate angle
            angle = calculate_angle(shoulder, elbow, wrist)
            
            # Visualize angle
            cv2.putText(image, str(angle), 
                           tuple(np.multiply(elbow, [640, 480]).astype(int)), 
                           cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255, 255, 255), 2, cv2.LINE_AA
                                )
            
            # Curl counter logic
 
          
            if angle > 30:
                stage = "sit"
                sit_count.append(1)

            if  stage =='sit' and angle < 15 :
                stage="stand"
                counter = time.time() - start_time 
                counter = float("{0:.2f}".format(counter))
                print(counter,"seconds")
                sit_to_stand = float("{0:.2f}".format(counter-time_spare))
                print("Sit to stand",sit_to_stand,"seconds")
                time_spare = counter 
                sit_count.append(0)
       
                
                       
        except:
            pass
        
        # Render curl counter
        # Setup status box
        cv2.rectangle(image, (0,0), (225,73), (245,117,16), -1)
        
        # Rep data
        cv2.putText(image, 'REPS', (15,12), 
                    cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0,0,0), 1, cv2.LINE_AA)
        cv2.putText(image, str(counter), 
                    (10,60), 
                    cv2.FONT_HERSHEY_SIMPLEX, 2, (255,255,255), 2, cv2.LINE_AA)
        
        # Stage data
        cv2.putText(image, 'STAGE', (65,12), 
                    cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0,0,0), 1, cv2.LINE_AA)
        cv2.putText(image, stage, 
                    (60,60), 
                    cv2.FONT_HERSHEY_SIMPLEX, 2, (255,255,255), 2, cv2.LINE_AA)
        
        
        # Render detections
        mp_drawing.draw_landmarks(image, results.pose_landmarks, mp_pose.POSE_CONNECTIONS,
                                mp_drawing.DrawingSpec(color=(245,117,66), thickness=2, circle_radius=2), 
                                mp_drawing.DrawingSpec(color=(245,66,230), thickness=2, circle_radius=2) 
                                 )               
        out.write(image)
        cv2.imshow('Mediapipe Feed', image)

        if cv2.waitKey(10) & 0xFF == ord('q'):
            total_time = float("{0:.2f}".format(time.time() - start_time))
            print("Total",total_time,"seconds")
            break
        elif sit_count[-2:] == [0,1]:
            total_time = float("{0:.2f}".format(time.time() - start_time))
            print("Total",total_time,"seconds")
           
            cap.release()
            out.release()
            break
           

    cap.release()
    cv2.destroyAllWindows()
# append dataframe
df =  df.append({"No_Name": run_no
                    , 'Name':f'output_{str(now)}_run_no{run_no}.mp4'
                    , 'Sit Total': sit_to_stand
                    , 'Total': str(now)
                    , 'Time': total_time}, ignore_index=True)
print("run complete")

Main code

#open camera
cap = cv2.VideoCapture(0)
# Curl counter variables
counter = 0
stage = None
i=0
start_time = time.time()
time_spare = 0
test = True
prev_frame_time = 0
new_frame_time = 0
sit_count = []

## Setup mediapipe instance
with mp_pose.Pose(min_detection_confidence=0.5, min_tracking_confidence=0.5) as pose:
    while cap.isOpened():
        ret, frame = cap.read()
        # Recolor image to RGB
        image = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
        image.flags.writeable = False
        # Make detection
        results = pose.process(image)
        # Recolor back to BGR
        image.flags.writeable = True
        image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
        image_height, image_width, _ =  image.shape
        image = cv2.resize(image, (int(image_width * (480 / image_height)), 480))
        # Extract landmarks
        try:
            landmarks = results.pose_landmarks.landmark
            # Get coordinates
            shoulder = [landmarks[mp_pose.PoseLandmark.LEFT_HIP.value].x,landmarks[mp_pose.PoseLandmark.LEFT_HIP.value].y]
            elbow = [landmarks[mp_pose.PoseLandmark.LEFT_KNEE.value].x,landmarks[mp_pose.PoseLandmark.LEFT_KNEE.value].y]
            wrist = [landmarks[mp_pose.PoseLandmark.LEFT_WRIST.value].x,landmarks[mp_pose.PoseLandmark.LEFT_WRIST.value].y]
            # Calculate angle
            angle = calculate_angle(shoulder, elbow, wrist)
            # Visualize angle
            cv2.putText(image, str(angle), 
                           tuple(np.multiply(elbow, [640, 480]).astype(int)), 
                           cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255, 255, 255), 2, cv2.LINE_AA
                                )
            # Curl counter logic
          
            if angle > 30:
                stage = "sit"
                sit_count.append(1)
            if  stage =='sit' and angle < 15 :
                stage="stand"
                counter = time.time() - start_time
                counter = float("{0:.2f}".format(counter))
                print(counter,"seconds")
                print("Sit to stand",float("{0:.2f}".format(counter-time_spare)),"seconds")
                time_spare = counter
                sit_count.append(0)  
                
        except:
            pass
        # Render curl counter
        # Setup status box
        cv2.rectangle(image, (0,0), (225,73), (245,117,16), -1)
        # Rep data
        cv2.putText(image, 'REPS', (15,12), 
                    cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0,0,0), 1, cv2.LINE_AA)
        cv2.putText(image, str(counter), 
                    (10,60), 
                    cv2.FONT_HERSHEY_SIMPLEX, 2, (255,255,255), 2, cv2.LINE_AA)
        # Stage data
        cv2.putText(image, 'STAGE', (65,12), 
                    cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0,0,0), 1, cv2.LINE_AA)
        cv2.putText(image, stage, 
                    (60,60), 
                    cv2.FONT_HERSHEY_SIMPLEX, 2, (255,255,255), 2, cv2.LINE_AA)
        # Render detections
        mp_drawing.draw_landmarks(image, results.pose_landmarks, mp_pose.POSE_CONNECTIONS,
                                mp_drawing.DrawingSpec(color=(245,117,66), thickness=2, circle_radius=2), 
                                mp_drawing.DrawingSpec(color=(245,66,230), thickness=2, circle_radius=2) 
                                 )               
        cv2.imshow('Mediapipe Feed', image)
        if cv2.waitKey(10) & 0xFF == ord('q'):
            print("Total",float("{0:.2f}".format(time.time() - start_time)),"seconds")
            break
        elif sit_count[-2:] == [0,1]:
            print("Total",float("{0:.2f}".format(time.time() - start_time)),"seconds")
            break
        
    cap.release()
    cv2.destroyAllWindows()

Example of the output

2
4.55 seconds
Sit to stand 4.55 seconds
Total 5.73 seconds
run complete

Send data to the google sheet