hand gesture data classification and recognize gesture from image

Hi everybody, i have a problem with my hand gesture recognition and i really need help. My system overview:
  • Detect hand from an image
  • Taking edge of the hand by Canny edge detection.
  • Feature extraction using Local contour sequence (LCS)
  • Classification.
I completed the 3rd step and got feature extraction from data sample images and put them into vectors. How do i use my data to train for my system and start recognize input image? I saw that somebody use LINEAR CLASSIFIER technique to recognize gesture, somebody use MULTI CLASS SUPPORT VECTOR MACHINE, but i don't know how to apply those techniques to train my system and recognize my gesture. Can anybody help me please?

7 件のコメント

Jan
Jan 2012 年 11 月 29 日
The message appeared 3 times. I've deleted the other message. What has happened?
Image Analyst
Image Analyst 2012 年 11 月 29 日
I think training you how to do that would go way way beyond what people have traditionally offered from this forum. It wouldn't be a half page answer or so. From the FAQ: http://matlab.wikia.com/wiki/FAQ?&cb=5719#Can_you_program_up_the_algorithm_in_this_article_for_me_and_explain_it_to_me.3F
Explorer
Explorer 2013 年 9 月 21 日
Hello nguyen. Can you please tell me how you detected hand from an image?
Image Analyst
Image Analyst 2013 年 9 月 21 日
He may have picked a published algorithm from section 21.4 in Vision Bib
romaisa sabih
romaisa sabih 2016 年 3 月 17 日
sir i also have to detect hand from image can you please share your code.Thanks in advance
Anuj Kumar
Anuj Kumar 2016 年 5 月 5 日
sir how did you find local contour sequence in matlab.Please share this code. Myemail id is anujmunger@gmail.com
NAVNEET NAYAN
NAVNEET NAYAN 2019 年 4 月 7 日
@ Image Analyst...When I am clicking the Viison Bib I am not geeting the path of section 21.4. Can you please tell me how to reach that section.

サインインしてコメントする。

 採用された回答

Kenta
Kenta 2020 年 3 月 29 日

0 投票

In this case, one solution is a use of CNN and LSTM for video (gesture) classification.
Although you might focus on classfication with a still image this time, a classification with a series of frames would be more powerful.
You may check the example below.

その他の回答 (1 件)

Ilya
Ilya 2012 年 11 月 29 日

0 投票

Perhaps you could start by reading the User Guide for the Statistics Toolbox, where many tools for classification are collected:

1 件のコメント

nguyen
nguyen 2012 年 11 月 29 日
i found that svmtrain function can train my system to recognize gesture. Suppose that i have 3 vectors: [1 0 1 1 1] [0 1 0 1 1] [1 0 0 1 1] these 3 vectors are results i collected after feature extraction, both of them represent for image 'A'. Can you tell me how to use svmtrain to train the system and which function do we use to input a new image and recognize it?

サインインしてコメントする。

質問済み:

2012 年 11 月 29 日

回答済み:

2020 年 3 月 29 日

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by