The variety methods for recognizing the action based on the feature of human skeleton.

GRU : (간단한 설명)

LSTM : Mediapipe로 추출한 스켈레톤 정보를 LSTM의 입력으로 사용하여 운동 동작을 분류합니다.

Posture : It only look one skeleton's coordinate and inference what is current action(posture).

SkeketonMHI_ObjectDetection : detect the Skeleton MHI and recognize the action

Transformer : (간단한 설명)

WEB: Web UI of main model, LSTM

It extracts the normalized coordinates by using Mediapipe.