線上教室

程式安全-姿勢辨識

講義

姿勢辨識

圖片

# @title Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# https://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

安裝套件

!pip install -q mediapipe

下載檔案

!wget -O pose_landmarker.task -q https://storage.googleapis.com/mediapipe-models/pose_landmarker/pose_landmarker_heavy/float16/1/pose_landmarker_heavy.task

Visualization utilities

# @markdown To better demonstrate the Pose Landmarker API, we have created a set of visualization tools that will be used in this colab. These will draw the landmarks on a detect person, as well as the expected connections between those markers.

from mediapipe import solutions
from mediapipe.framework.formats import landmark_pb2
import numpy as np


def draw_landmarks_on_image(rgb_image, detection_result):
  pose_landmarks_list = detection_result.pose_landmarks
  annotated_image = np.copy(rgb_image)

  # Loop through the detected poses to visualize.
  for idx in range(len(pose_landmarks_list)):
    pose_landmarks = pose_landmarks_list[idx]

    # Draw the pose landmarks.
    pose_landmarks_proto = landmark_pb2.NormalizedLandmarkList()
    pose_landmarks_proto.landmark.extend([
      landmark_pb2.NormalizedLandmark(x=landmark.x, y=landmark.y, z=landmark.z) for landmark in pose_landmarks
    ])
    solutions.drawing_utils.draw_landmarks(
      annotated_image,
      pose_landmarks_proto,
      solutions.pose.POSE_CONNECTIONS,
      solutions.drawing_styles.get_default_pose_landmarks_style())
  return annotated_image

實作1

import cv2
from google.colab.patches import cv2_imshow

img = cv2.imread("exercise.png")
cv2_imshow(img)

程式碼

# STEP 1: Import the necessary modules.
import mediapipe as mp
from mediapipe.tasks import python
from mediapipe.tasks.python import vision

# STEP 2: Create an PoseLandmarker object.
base_options = python.BaseOptions(model_asset_path='pose_landmarker.task')
options = vision.PoseLandmarkerOptions(
    base_options=base_options,
    output_segmentation_masks=True)
detector = vision.PoseLandmarker.create_from_options(options)

# STEP 3: Load the input image.
image = mp.Image.create_from_file("exercise.png")

# STEP 4: Detect pose landmarks from the input image.
detection_result = detector.detect(image)

# STEP 5: Process the detection result. In this case, visualize it.
annotated_image = draw_landmarks_on_image(image.numpy_view(), detection_result)
cv2_imshow(cv2.cvtColor(annotated_image, cv2.COLOR_RGB2BGR))

執行結果

Visualize the pose segmentation mask.

程式碼

segmentation_mask = detection_result.segmentation_masks[0].numpy_view()
visualized_mask = np.repeat(segmentation_mask[:, :, np.newaxis], 3, axis=2) * 255
cv2_imshow(visualized_mask)

執行結果

https://www.ithome.com.tw/news/126911

實作2

影片

下載影片檔案連結

將影片檔轉成圖片集

import cv2
import os

# Function to extract frames from a video until reaching the desired frame count
def extract_frames(video_file):
    cap = cv2.VideoCapture(video_file)

    frame_rate = 30  # Desired frame rate (1 frame every 0.5 seconds)
    frame_count = 0

    # Get the video file's name without extension
    video_name = os.path.splitext(os.path.basename(video_file))[0]

    # Create an output folder with a name corresponding to the video
    output_directory = f"{video_name}_frames"
    os.makedirs(output_directory, exist_ok=True)

    while True:
        ret, frame = cap.read()

        if not ret:
            break

        frame_count += 1

        # Only extract frames at the desired frame rate
        if frame_count % int(cap.get(5) / frame_rate) == 0:
            output_file = f"{output_directory}/frame_{frame_count}.jpg"
            cv2.imwrite(output_file, frame)
            print(f"Frame {frame_count} has been extracted and saved as {output_file}")

    cap.release()
    cv2.destroyAllWindows()

if __name__ == "__main__":
    video_file = r"motion01.mp4"  # Replace with your video's name

    extract_frames(video_file)

安裝套件

!pip install -q mediapipe

取得模型檔案

!wget -O pose_landmarker.task -q https://storage.googleapis.com/mediapipe-models/pose_landmarker/pose_landmarker_heavy/float16/1/pose_landmarker_heavy.task

對圖片集進行辨識

import cv2
from google.colab.patches import cv2_imshow
# STEP 1: Import the necessary modules.
import mediapipe as mp
from mediapipe.tasks import python
from mediapipe.tasks.python import vision

# STEP 2: Create an PoseLandmarker object.
base_options = python.BaseOptions(model_asset_path='pose_landmarker.task')
options = vision.PoseLandmarkerOptions(
    base_options=base_options,
    output_segmentation_masks=True)
detector = vision.PoseLandmarker.create_from_options(options)
os.mkdir("motion01_detect")

for i in range(84):
  # Use the file path directly instead of reading it with cv2.imread
  image_path = f"/content/motion01_frames/frame_{i+1}.jpg" 
  
  # STEP 3: Load the input image using the file path.
  image = mp.Image.create_from_file(image_path)  

  # STEP 4: Detect pose landmarks from the input image.
  detection_result = detector.detect(image)

  # STEP 5: Process the detection result. In this case, visualize it.
  annotated_image = draw_landmarks_on_image(image.numpy_view(), detection_result)
  # cv2_imshow(cv2.cvtColor(annotated_image, cv2.COLOR_RGB2BGR))

  # 寫入圖檔
  cv2.imwrite(f"/content/motion01_detect/frame_{i+1}.jpg", annotated_image)

相關函式

# @markdown To better demonstrate the Pose Landmarker API, we have created a set of visualization tools that will be used in this colab. These will draw the landmarks on a detect person, as well as the expected connections between those markers.

from mediapipe import solutions
from mediapipe.framework.formats import landmark_pb2
import numpy as np


def draw_landmarks_on_image(rgb_image, detection_result):
  pose_landmarks_list = detection_result.pose_landmarks
  annotated_image = np.copy(rgb_image)

  # Loop through the detected poses to visualize.
  for idx in range(len(pose_landmarks_list)):
    pose_landmarks = pose_landmarks_list[idx]

    # Draw the pose landmarks.
    pose_landmarks_proto = landmark_pb2.NormalizedLandmarkList()
    pose_landmarks_proto.landmark.extend([
      landmark_pb2.NormalizedLandmark(x=landmark.x, y=landmark.y, z=landmark.z) for landmark in pose_landmarks
    ])
    solutions.drawing_utils.draw_landmarks(
      annotated_image,
      pose_landmarks_proto,
      solutions.pose.POSE_CONNECTIONS,
      solutions.drawing_styles.get_default_pose_landmarks_style())
  return annotated_image

將從影片擷取出的圖片進行辨識

import cv2
from google.colab.patches import cv2_imshow
# STEP 1: Import the necessary modules.
import mediapipe as mp
from mediapipe.tasks import python
from mediapipe.tasks.python import vision

# STEP 2: Create an PoseLandmarker object.
base_options = python.BaseOptions(model_asset_path='pose_landmarker.task')
options = vision.PoseLandmarkerOptions(
    base_options=base_options,
    output_segmentation_masks=True)
detector = vision.PoseLandmarker.create_from_options(options)
os.mkdir("motion01_detect")

for i in range(84):
  # Use the file path directly instead of reading it with cv2.imread
  image_path = f"/content/motion01_frames/frame_{i+1}.jpg"

  # STEP 3: Load the input image using the file path.
  image = mp.Image.create_from_file(image_path)

  # STEP 4: Detect pose landmarks from the input image.
  detection_result = detector.detect(image)

  # STEP 5: Process the detection result. In this case, visualize it.
  annotated_image = draw_landmarks_on_image(image.numpy_view(), detection_result)
  # cv2_imshow(cv2.cvtColor(annotated_image, cv2.COLOR_RGB2BGR))

  # 寫入圖檔
  cv2.imwrite(f"/content/motion01_detect/frame_{i+1}.jpg", annotated_image)

將辨識後的圖片會整成影片

import cv2

# 表示 MP4 編碼格示，副檔名為 .mp4
fourcc = cv2.VideoWriter_fourcc('M', 'P', '4', 'V')

out = cv2.VideoWriter("motion01_detect.mp4", fourcc, 1, (1280,720))

for i in range(84):
  # Use the file path directly instead of reading it with cv2.imread
  image_path = f"/content/motion01_detect/frame_{i+1}.jpg"

  frame = cv2.imread(image_path)
  out.write(frame)

out.release()
cv2.destroyAllWindows()