frame by frame from a USB capture device

Accelerated Computing CUDA CUDA Programming and Performance
1

Hi,

How do I go about processing video frame by frame from a USB video capture device? I can’t seem to piece together how to do it from all the examples I gone through.

I think its something like

USB driver -> windows DirectDraw -> CUDA

But I really have no idea, and I’d like to cut out the middle step if possible?

Sorry for such a basic question, but this is my first week with programming for Windows and CUDA and I can’t seem to work out how to do it.

Any help is appreciated.

2

Just a little point in the right direction would really help

3

Hi JobyBett,
Would you please explain more about what you are doing?

4

The end goal is to take a video stream, in my case a feed from a cable box, using a Video to USB capture device, then parse the video frames looking for objects, say detect a menu has appeared on the screen, then OCR the text to see what menu it is. I want to do this in real time, but probably only process 10 frames in every second.

The OCR is well down the line, right now I’m just struggling with working out how to get the video into the CUDA environment and process it frame by frame.

I’ve taken a look at the Sobel filter sample, and understand how to do things like edge detection, but the sample comes from a static .pgm file

5

I must be going in the wrong direction completely, I still have not been able to find any information on this,

All I want to do is feed the video captured by a USB capture device into my CUDA program so that I can process it frame by frame.

6

did you try the openCV examples yet?