Modeling - [Process] 스윙 결과 비교 설계

[UGeunJi]

고민점: 스윙 결과 별로 landmark의 좌표를 통해 각도와 위치로 일치율을 판단해야 한다.

• 2차원일 때

landmark의 좌표는 x, y로 2차원으로 나타나지만 이미지의 전체 크기에 따라 상대적으로 값이 변한다.

• 3차원일 때

landmark의 좌표가 x, y, z로 3차원으로 나타난다. 이미지 크기와 상관없이 skeleton의 크기에 의해 자동으로 resize된다.

Dimension	장점	단점
2차원	좌표값을 바로 응용할 수 있음	이미지 크기에 따라 상대적으로 좌표값이 바뀌는 것을 계산해야 한다.
3차원	100% 정확한 값을 얻을 수 있다.	3차원의 값을 가지고 2차원 이미지에 표시해줘야 한다.

---

방안

• 입력 이미지의 사람 크기 통일 (bounding box 이용)
• 3차원 방법

---

코드 (Colab)

• mediapipe download

!pip install mediapipe

• file 선택

from google.colab import files
uploaded = files.upload()

• 선택한 file 시각화

import cv2
from google.colab.patches import cv2_imshow
import math
import numpy as np

DESIRED_HEIGHT = 480
DESIRED_WIDTH = 480
def resize_and_show(image):
  h, w = image.shape[:2]
  if h < w:
    img = cv2.resize(image, (DESIRED_WIDTH, math.floor(h/(w/DESIRED_WIDTH))))
  else:
    img = cv2.resize(image, (math.floor(w/(h/DESIRED_HEIGHT)), DESIRED_HEIGHT))
  cv2_imshow(img)

# Read images with OpenCV.
images = {name: cv2.imread(name) for name in uploaded.keys()}
# Preview the images.
for name, image in images.items():
  print(name)   
  resize_and_show(image)

import mediapipe as mp
mp_pose = mp.solutions.pose
mp_drawing = mp.solutions.drawing_utils 
mp_drawing_styles = mp.solutions.drawing_styles

help(mp_pose.Pose)

• landmark 2차원 좌표값 계산 후 시각화

# Run MediaPipe Pose and draw pose landmarks.
with mp_pose.Pose(
    static_image_mode=True, min_detection_confidence=0.5, model_complexity=2) as pose:
  for name, image in images.items():
    # Convert the BGR image to RGB and process it with MediaPipe Pose.
    results = pose.process(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))
    
    # Print nose landmark.
    image_hight, image_width, _ = image.shape
    if not results.pose_landmarks:
      continue
    print(
      f'Nose coordinates: ('
      f'{results.pose_landmarks.landmark[mp_pose.PoseLandmark.NOSE].x * image_width}, '
      f'{results.pose_landmarks.landmark[mp_pose.PoseLandmark.NOSE].y * image_hight})'
    )
    print(
      f'L_elbow coordinates: ('
      f'{results.pose_landmarks.landmark[mp_pose.PoseLandmark.LEFT_ELBOW].x * image_width}, '
      f'{results.pose_landmarks.landmark[mp_pose.PoseLandmark.LEFT_ELBOW].y * image_hight})'
    )
    print(
      f'R_elbow coordinates: ('
      f'{results.pose_landmarks.landmark[mp_pose.PoseLandmark.RIGHT_ELBOW].x * image_width}, '
      f'{results.pose_landmarks.landmark[mp_pose.PoseLandmark.RIGHT_ELBOW].y * image_hight})'
    )
    print(
      f'L_hip coordinates: ('
      f'{results.pose_landmarks.landmark[mp_pose.PoseLandmark.LEFT_HIP].x * image_width}, '
      f'{results.pose_landmarks.landmark[mp_pose.PoseLandmark.LEFT_HIP].y * image_hight})'
    )
    print(
      f'R_hip coordinates: ('
      f'{results.pose_landmarks.landmark[mp_pose.PoseLandmark.RIGHT_HIP].x * image_width}, '
      f'{results.pose_landmarks.landmark[mp_pose.PoseLandmark.RIGHT_HIP].y * image_hight})'
    )
    print(
      f'L_foot_index coordinates: ('
      f'{results.pose_landmarks.landmark[mp_pose.PoseLandmark.LEFT_FOOT_INDEX].x * image_width}, '
      f'{results.pose_landmarks.landmark[mp_pose.PoseLandmark.LEFT_FOOT_INDEX].y * image_hight})'
    )
    print(
      f'R_foot_index coordinates: ('
      f'{results.pose_landmarks.landmark[mp_pose.PoseLandmark.RIGHT_FOOT_INDEX].x * image_width}, '
      f'{results.pose_landmarks.landmark[mp_pose.PoseLandmark.RIGHT_FOOT_INDEX].y * image_hight})'
    )

    # Draw pose landmarks.
    print(f'Pose landmarks of {name}:')
    annotated_image = image.copy()
    mp_drawing.draw_landmarks(
        annotated_image,
        results.pose_landmarks,
        mp_pose.POSE_CONNECTIONS,
        landmark_drawing_spec=mp_drawing_styles.get_default_pose_landmarks_style())
    resize_and_show(annotated_image)

• 각도 계산

import math

def get_num():
  x1 = results.pose_landmarks.landmark[mp_pose.PoseLandmark.RIGHT_HIP].x * image_width
  y1 = results.pose_landmarks.landmark[mp_pose.PoseLandmark.RIGHT_HIP].y * image_hight
  x2 = results.pose_landmarks.landmark[mp_pose.PoseLandmark.LEFT_FOOT_INDEX].x * image_width
  y2 = results.pose_landmarks.landmark[mp_pose.PoseLandmark.LEFT_FOOT_INDEX].y * image_hight
  return [x1, y1, x2, y2]

def cal_rad(arr):
  rad = math.atan2(arr[3]-arr[1],arr[2]-arr[0])
  return rad

def radTodeg(rad):
  PI = math.pi
  deg = (rad*180)/PI
  print("{0:0.2f}".format(deg))

arr = get_num()
rad = cal_rad(arr)
radTodeg(rad)

• landmark 3차원 계산 좌표

# Run MediaPipe Pose and plot 3d pose world landmarks.
with mp_pose.Pose(
    static_image_mode=True, min_detection_confidence=0.5, model_complexity=2) as pose:
  for name, image in images.items():
    results = pose.process(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))

    # Print the real-world 3D coordinates of nose in meters with the origin at
    # the center between hips.
    print('Nose world landmark:'),
    print(results.pose_world_landmarks.landmark[mp_pose.PoseLandmark.NOSE])
    
    # Plot pose world landmarks.
    mp_drawing.plot_landmarks(
        results.pose_world_landmarks, mp_pose.POSE_CONNECTIONS)