TipTep Skeletonizer is an open source middleware, based on OpenNI platform, which provides tools and APIs that can be used to extract the geometric skeleton of the hand from depth images. Geometric skeleton of the hand is a simplified version of hand`s depth image, which contains all necessary information about hand`s gesture and can be used to store and compare hand/finger gestures.

Introduction


Some types of HCI applications require to store hand`s gesture information in the database, compare gestures and analyze hand`s posture. To solve these tasks it is inefficient to use hand`s whole depth or RGB images, because:

  1. Images require big space in the database

  2. It is time costly to compare images

  3. Image analysis require knowledge of computer vision algorithms

As an official middleware partner of OpenNI consortium we provide a library, which makes these tasks easier. In computer vision it is a popular technique to use object`s geometric skeleton to store and compare two objects in the image. Now you can use this technique to analyze and recognize hand`s custom gestures via hand`s image simplified descriptor – hand`s skeleton. 

Resources