A Gesture Recognition App using Microsoft Kinect V2
This is a windows desktop application for gesture recognition and their mapping to various mouse and keyboard inputs.
The WPF application aims at controlling mouse and keyboard inputs with innovatively mapped hand gestures(both continuous and discrete).
The application authenticates and also dynamically authorizes users to use gestures for interactive control over the system.
The intuitive UI design helps users set sensitivity and allows them to create authorization profiles dynamically.
Kinect is a motion sensing input device by Microsoft for the Xbox 360 video game console and Windows
PCs. Based around a webcam-style add-on peripheral for the Xbox 360 console, it enables users to control and
interact with the Xbox 360 without the need to touch a game controller, through a natural user interface using gestures
and spoken commands. Kinect Studio is a tool provided by Microsoft Kinect SDK 2.0 to record the gestures.
This recorded gesture is a raw .xrf file which then has to be converted into an executable .xef clip. The gesture is then
trained for true and false values using Visual Gesture Builder. The trained gesture is then stored in a database which
can be imported in a Visual Studio Project to bind actions to this gesture.
Microsoft Kinect V2 comes with powerful infrared sensers and depth sensers which can detect upto 6 skeletal frames at once.
However, the challenge here is to identify the authorized users in real-time.
This can be achieved with face recognition. For this we use dimensionality reduction using PCA(Principal Component Analysis) and Eigen Faces[1].
These are low level features for grey scale images of faces that then can be compared with newly captured images at real time.
Authorized users will be allowed to use gestures to control the system. New users can be added and authorized dynamically using the same tool.
The tool comes with two sections: Settings and Authentication.
Gestures are built using Microsoft Kinect Visual Gesture Builder tool provided by Microsoft.
The research paper was published in International Journal of Innovative Research in Computer and Communication Engineering in the year 2016, volume 4, issue 5. Checkout the paper here
Chethan M R Adarsh M V Akarsh M S Karunesh Gurkar,Prashanth Kumar(Guide)(2016).“Gesture Recognition for Interactive Sys-tems Using Kinect v2”. In:International Journal of Innovative Re-search in Computer and Communication Engineering(IJIRCCE),4(5):9186–9191,052016.ISSN:2320-9801. DOI:10.15680/IJIRCCE.2015.0405108. URL:http://www.ijircce.com/upload/2016/may/108_22_Gesture_2%20HARD.pdf
The project was awarded 3rd place in the Annual Project Exhibition, “Vivaceous”.