/Fall_Detection_Project

此專案是針對一連串人體的骨幹座標(frames)變化來偵測人體的目前動作的 CNN，訓練資料來自 SBU_dataset 以及我自行利用拍攝好的照片經過 openpose 得出人體骨幹座標的資料

Primary LanguagePython

Fall Detection with Co-occurrence-Feature-Learning-From-Skeleton-2D-Data

此專案是針對一連串人體的骨幹座標(frames)變化來偵測人體的目前動作的 CNN，訓練資料來自 SBU_dataset 以及我自行利用拍攝好的照片經過 openpose 得出人體骨幹座標的資料

Detail

-This is a project implemented with Keras based on the code from here, relative paper.

-Training and Testing data is in data_seq dirctory.

-The whole model is defined in model.py.

-Addtional data-processing code is in utils.py.

-To test the model, use fall_detection.py.

且部分資料由熱成像圖轉換成座標圖來做訓練，如下 :

class 9

fall_img fall_img2

class 10

sit_img sit_img2

Result

Train with 416 skeleton sequences (20 frames)

Test with 83 skeleton sequences (20 frames)

Training acc : 100 %

Testing acc : 94 %