yoga pose detection correction
1.0.0
優選圖像應為JPG/JPEG,圖像名稱應為[number].jpg 。
./poses_dataset/Images中創建一個新目錄(名稱可以是任何東西,但我建議使用姿勢的名稱),並用姿勢圖像填充它。./poses_dataset/angles中創建另一個目錄(文件夾名稱應與步驟1中使用的名稱相同),並放置一個姿勢的圖像。該目錄中的圖像將用作“已知的好”姿勢角度(姿勢應該是完美的),如在實時檢測過程中,將用戶的姿勢與此姿勢進行比較以提出建議。create_poses_csv.ipynb 。 This will create a file named https://github.com/bourbonbourbon/yoga-pose-detection-correction/blob/main/create_poses_csv.ipynb?short_path=1#L120 (you can name it whatever) which has all the x, y, z, and visibility values of all the desired landmark points of all poses in the ./poses_dataset/Images目錄。生成的CSV文件中的姿勢列值將是整數。create_angles_csv.ipynb 。這將創建另一個名為https://github.com/bourbonbourbon/yoga-pose-detection-detection-correction/blob/main/main/create_angles_csv.ipynb?short_path=1#l104的CSV。rfc_model.ipynb ,該rfc_model.ipynb使用步驟3中生成的CSV作為輸入文件來訓練/測試數據。然後,它將創建一個.model文件,名為https://github.com/bourbonbourbon/yoga-pose-detection-c-c- c-correction/blob/main/main/rfc_model.ipynb?short_short_short_short_path=1#l44live_detection.py https://github.com/bourbonbourbon/yoga-pose-detection-correction/blob/main/live_detection.py#L106 https://github.com/bourbonbourbon/yoga-pose-detection-correction/blob/main/live_detection.py#L108 to whatever you have created in steps 4和5。注意:請參閱https://github.com/bourbonbourbon/yoga-pose-detection-correction/tree/1c9a4e50c00be9a8b677632901e6b8b0c459b6f4用於項目結構。
pip install -r requirements.txt安裝來自requirements.txt的所有庫.txt。python live_detection.py 。