Openni is a simple and efficient open-source framework developed by Kinect. It is easy to use. Many universities and research institutions choose openni as a development platform instead of a powerful Kinect SDK, here we will introduce several Middleware Based on openni:
1. nite2:
Complete gesture control and body control APIs, gesture locating and tracking, and human skeleton tracking ......
Advantages: Low CPU usage, simple and flexible, multi-platform support (Windows, Linux, Mac OS and andorid)
URL: http://www.primesense.com/solutions/nite-middleware/
2. 3D Hand Tracking Library
Provides a complete 3D gesture tracking framework API
Advantage: you do not need to mark the position, direction, and joint of 3D gestures.
Disadvantage: the platform has high requirements,
Presequisites: windows7 64bit or Ubuntu 12.04 64bit workstation. nvidia gpu withcuda support.
URL: http://cvrlcode.ics.forth.gr/handtracking/
3. sigmanil framework
Provides powerful natural user interaction excuses, including: gesture reconstruction (to achieve finger precision), gesture recognition, and gesture skeleton tracking.
Advantages: fully open-source and complete framework
Presequisites: openni and Nite or kinectsdk. Some modules and samples need boost, opencv, Cuda toolkit, and the sample requires QT.
URL: http://www.sigmanil.com
4. tiptep skeletonizer
Provides the geometric skeleton of a gesture Based on the depth of the gesture.
Presequisites: Windows x86/x64,. Net 4.
URL: http://tiptep.com/index.php/products
5. 3D Face Identification
Principle: 1. use opencv face detection to locate faces in RGB Images; 2. convert the depth data of the face area into well-defined depth data in advance. 3. compare the detected data with the data in the database.
Advantage: The program provides basic function sets for detection and recognition for secondary development. Real-time.
Reference: [1] r. min, J. choi, G. medioni, J.-L. dugelay, "real-time 3D faceidentification from a depth camera", 21st International Conference on patternrecognition (ICPR), November 11-15,201 2.
Presequisites: GPU-programmed video card.
URL:
Http://iris.usc.edu/USC-Computer-Vision.html
6. volumental
Volumental is the fastest and easy way to scan and share any object you think of including a person, object, or room. Volumental provides a plug-in for your browser to connect to your deep camera, and uses specialized hardware and algorithms to recreate 3D models on the cloud.
Presequisites: Windows platform. verified to work with Safari, Firefox and chrome
URL: http://www.volumental.com/
7. kscan3d Middleware
Kscan3d provides a complete solution for extracting and processing 3D scan data, including a robust API and LUA-Based Automated scripts. The integrated Script Editor provides a syntax for outputting and debugging scripts based on the logs in the GUI console. You can also use the API
DLL provides functions from kscan3d for your project, including scanning, alignment, and grid processing.
Presequisites: Windows 7 or8 (64-bit recommended)
URL:
Http://www.kscan3d.com/
Kscan3d and SDK 1.7's Kinect
Fusion functions are similar, but it can only be used for 14 days for free.
8. viim SDK v2.0.0 Beta
Features are simplified based on openni and Nite to add high-level features. Viim SDK aims to provide more general Nui for every programmer. It is actually an enhanced version of openni/nite.
Presequisites: Windows 7 orlater; x64/x86;
URL:
Http://www.viim.pt/
9. GST API
GST (globalsensing technologies) provides a very simple program to identify projects, and gstskeletonpostures can easily mark user actions.
Presequisites: Windows 7 32or 64 bits, the last corresponding version of openn2and nite2 librariesinstalled. Project have been made using Microsoft SDK 7.1 and Microsoft VisualC ++ express 2010.
URL: http://globalsensing.eu/
10. Motion nexus plugin
Motion nexus in uses openni to provide Flash developers with tools and environments to build real immersive and interactive games and applications.
URL: http://www.motionnexus.com/openni.html
11. starry night feature extractiondemonstration SDK
The Feature Extraction SDK allows you to provide (1) a model of the expected object, and (2) a 3D scan of the scene containing the object. The software identifies and locates objects in 3D scanning scenarios. Van Gogh's star night proves the feature extraction capability of This SDK. It is only one of the many functions provided by the real-time 3d object recognition and analysis software.
Presequisites: PC running Windows 7 with openni 2.0 installed, Carmine Sensor
URL: http://www.vangoghimaging.com/3d _scanning_imaging_products.html
These programs are actually open-source. You can go to http://www.openni.org/to understand them ~