增强视觉 | 计算机视觉 增强现实

看上去很美。来自MIT，利用深度摄像头和触感作为输入设备，控制一个可上下浮动的一组实体键。

只不过，用来做什么呢？弹钢琴？变魔术么？说起弹钢琴，最近看到一个用kinect做输入，玩虚拟弦乐的东西，链接在此。

阅读全文…

俺看上这一篇了，各种期待。

http://research.microsoft.com/apps/pubs/default.aspx?id=145347

不解释

微软研究院在UIST上展示一个异想天开的UI技术：light space。思路是利用多个深度摄像头回复场景内的3D结构，然后利用人体完成一些奇怪搞笑的UI操作。比如，先左手接触一下A屏幕，再右手接触一下B屏幕，就可以A屏幕上的内容转移到B屏幕上。再比如用一个扫除的动作，可以把桌面上显示的内容”扫”到手上，并以图标的形式显示。。。

阅读全文…

之前CVChina介绍过几家TOF相机厂商,这里有一段PMD的实拍视频.

cvchina曾经介绍过几款深度摄像头，看这里，但是他们的价格都让人望而却步，不是谁都能承担得起几千美元的设备的。最近得知另一个消费者级别的深度摄像头（感谢某巨专业人士），价格应该在200美元左右，跟快要上市的kinect差不多一个价位。

阅读全文…

如此多的手势UI，德国Fraunhofer的手势UI也来插一脚（插一手？），特色是可以在3D空间操作的哈。

个人觉得一个好的手势UI最重要的地方就是能不能提供一个简洁的抓取动作（相当于鼠标的左键点击）。在下面的视频里，Fraunhofer的抓取动作就是直接取自自然而然的手掌抓取，前提当然是建立在五个手指准确的的姿态估计之上。

另：在来源处提到，手势的3D信息来源于3D摄像头。

Im FIT-Prototyp werden in Echtzeit Hände und Finger der Benutzter in den Bilddaten einer 3D-Kamera erkannt und die Bewegungen mitverfolgt.

几段深度摄像机TOF的视频

阅读全文…

随着微软kinect(natal)的发展，深度摄像头吸引越来越多人的目光，深度摄像头可以用在人体跟踪，三维重建，人机交互，SLAM等等领域。但是深度摄像头的高昂的价格实在是让一般人望而却步，我所知道的primesense
的一个摄像头要5000美元。。。而kinect的出现会不会带动民用（相对廉价）的深度摄像头的发展呢（传闻kinect定价199美元）？
cvchina曾经介绍过两种深度摄像头，下面再转一个最近看到的几种深度摄像头的简介：

阅读全文…

介绍两个深度摄像头，第一个是3DV systems的ZCam，第二个是PrimeSense公司的PrimeSensor。两个公司都是以色列的，他们的摄像头的性能也非常相近，都可以达到QVGA分辨率下60FPS，RGB image 和 Depth Image同步并逐点对应。但连个摄像头的结构大不相同，而且所使用的技术的名称也不一样，一个是TIME OF FLIGHT， 一个是LIGHT CODING.
1.ZCam
ZCam is a brand of time-of-flight camera products for video applications by Israeli developer 3DV Systems. The ZCam supplements full-color video camera imaging with real-time range imaging information, allowing for the capture of video in 3D.

The original ZCam,[fn 1] released in 2000,[4] was an ENG video camera add-on used for digital video compositing.[5][6] Before agreeing in March 2009[7] to sell its assets to Microsoft,[8] 3DV had planned to release a ranging video webcam (previously called the Z-Sense), also under the name ZCam.[fn 1] The ZCam webcam was one of several competing real-time range imaging camera products in development that target home game controller applications.[fn 2]
———-摘自维基百科

2.PrimeSensor
PrimeSense is a fabless semiconductor company. Our technology empowers consumer electronic devices, such as TVs, set-top boxes, living-room PCs and more with natural interaction capabilities.
Our product, the PrimeSensor, contains the Reference Design and the NITE middleware.

The PrimeSensor Reference Design is a low-cost, plug and play, USB-powered device that can either sit on top of or next to a television screen or a monitor, or be integrated into them. The Reference Design generates realtime depth, color and audio data of the living room scene. It works in all room lighting conditions (whether in complete darkness or in a fully lit room). It does not require the user to wear or hold anything, does not require calibration and does not require computational resources from the host’s processor.
———摘自PrimeSense

natal

看到一则关于微软Natal的小道消息，全文照抄如下：

Yesterday, I was invited to a private Project Natal sneak peek. The event was held at the beautiful EZ Studios in New York City. I met with a member of the Project Natal product team. She gave us a brief explanation of the Natal vision and gave us only a few technical details:

• RBG camera
• depth camera
• Current games will not be compatible with Natal
• “Project Natal” is a code name and will not be the final name of the product

• [Work in progress] Use voice commands to control services i.e. music, videos, movies
• [Work in progress] Facial recognition
• [Work in progress] Use gesturing to navigate the XBOX menu structure
• [Work in progress] Distinguishing between the primary player and a spectator
• [Work in progress] Suppressing background noise

Once we were briefed — we were given the opportunity to play a dodge ball (like) game. It was a great experience. There I was (without a controller) standing, kicking, swatting, jumping and working up a sweat. The responsiveness of the Natal system was incredible and very accurate. If I physically moved forward, back, swung faster or slower my dodge ball avatar responded in kind.

“Our advances in computer vision and audio-signal processing,” Malvar notes, “have enabled the development of Project Natal, a new level of experience for Xbox, in which user gestures and voice take the place of the standard game controller.– Rico Malvar, managing director of Microsoft Research Redmond

语音识别可以多国语言么？可以自定义么？

来源