Skip to content

Latest commit

 

History

History
28 lines (27 loc) · 2.2 KB

README.md

File metadata and controls

28 lines (27 loc) · 2.2 KB

PredictSignLanguage_DNN

Sign language recognition using LeapMotion, DNN

Leapmotion을 활용하여 손동작에 대한 입력 값을 얻고 그 입력값을 학습하여 만든 DNN 모델

1. 립모션을 이용하여 손가락 동작에 대한 반환값을 전처리한다.

Leapmotion을 활용하였을때, 반환되는 값은 총 192개 (반환되는 값은 립모션 공식 홈페이지 참조)

2. 전처리한 결과값을 Python pandas module을 이용하여 csv 파일로 저장한다

Leapmotion에서 제공하는 sample.py 파일에 내가 필요한 값들을 전처리하여 엑셀 파일로 저장

3. csv 파일을 학습시킨다.

tensorflow 모듈을 이용

4. middle_present.py 파일은 학습한 모델을 불러와서 입력한 손동작이 어떠한 손동작인지 알려주는 파일

또 middle_present.py 에서 각 손가락에 대한 x, y, z 값에 대하여 곱하기 100을 함
값이 아주 작게 차이 나기 때문에 각 값에 대해 *100을 하여 값이 차이 많이 나도록 한다.

문의사항이 있으면, 이메일로 연락주시면 답변해드리겠습니다 liw5589@naver.com

DNN model that obtains input value for hand gesture by using Leapmotion and learns input value

1. Use Leapmotion to preprocess the return value for the finger gesture.

When using Leapmotion, the total value returned is 192 (refer to the lip motion official website for the returned value)

2. Save the preprocessed result as csv file using Python pandas module

Pre-process the values in the sample.py file provided by Leapmotion and save it as an Excel file

3. Train the csv file.

Using Tensorflow Module

4. The middle_present.py file is a file that imports the trained model and tells what type of hand gestures are entered.

If you have any questions, please contact us by email and we will answer. liw5589@naver.com