|
CSE Home |
About Us |
Search |
Contact Info |
Overview:
We frequently have speakers who either don't use Powerpoint, or can't provide the Powerpoint file to us long enough in advance of the talk to allow images to be prepared for inclusion in a media stream. For this scenario, we use a scan converter or another camera pointed at the screen to get an isolated video feed containing only the supporting materials. We then capture frames from this feed at an interval of approximately 10 seconds. The Image Capture Client tool described here runs on the system doing the frame captures. It detects new frame images, renames them and moves them to a web accessible driectory, constructs URL script commands corresponding to each new image, optionally manages a connection to a Script Command Server and sends script commands to the server for inclusion in a live media stream.
| Download runtime and source |
Capturing frames:
The system that captures frames from the video feed requires some special hardware and software. There are many possible candidates in this arena. The system we use is equipped with a Winnov Videum VO video capture card. The Videum software that comes with the card includes a "Web Cam" utility which is capable of capturing jpegs at a user-defined frequency. Other free and shareware web cam tools are available that may work with other capture hardware. There is source available for the Microsoft demo application 'vidcap32' which we have successfuly used in concert with "cjpeg" (a freeware jpeg compression source) to implement a crude "Video for Windows" compatible web cam.
Setup and Configuration:
Unzip the download above on the frame capture system. There is a setup program in the Runtime directory. If the system has a recent version of Visual Basic installed, it is probably not necessary to run setup. If in doubt, it won't hurt to run it. The executable is called "ImageCapClient.exe". Setup creates a shortcut to it in the Start menu. When setup is complete, run the Image Capture Client, and click "Configure".
If you are using the client in conjunction with a Script Command Server to feed URLs to a live stream, type the name or address of the server/encoder system, and the port number configured in the server. Otherwise, you can run in stand-alone mode. This will simply collect images in the destination directory, and write a log file.
Edit the Image Source Directory to be the local place where your frame capture software creates image files.
Edit the Base URL and Image Destination Directory to specify a place on your web server to where the images will be copied. The Base URL will be prepended to file name strings such as "image59.jpg" which will be constructed as the client detects and copies images. The Base URL and Image Destination Directory obviously need to refer to the same file space. We use a drive mapping to make the destination directory appear as a local drive.
Edit the local path for log files. Each time the client is launched, it will create a time-stamped log file in this directory. The log file contains all the image name and timing data which would be useful if you want to build the URLs into a Windows Media archive.
Operation:
As soon as the client launches it will attempt to establish and maintain a connection to a Script Command Server. The status of that connection is reported in the top line of its main form.
The order in which things are started is not important, but you ultimately should be running the frame capture software, and the Image Capture client on one system and optionally a Script Command Server and Windows Media Encoder on another system.
It is important that the source and destination directories exist before you click "Start" in the Image Capture Client. If Jpeg image files already exist in the source directory, the client will offer to delete them for you. If the destination directory is not empty, some checking is done to make sure there will not be name conflicts as the client runs.
After the client is started, it will check for new images in the source directory once every two seconds. If an image is found, it will be renamed "imageN.jpg" (where N is a sequence number), and moved to the destination directory. The corresponding script command will be composed using the specified base URL, then the full URL will be logged and sent to a connected Script Command Server.
|
Department of Computer Science & Engineering University of Washington Box 352350 Seattle, WA 98195-2350 (206) 543-1695 voice, (206) 543-2969 FAX [comments to E-mail the page owner] | |