Updating picture/canvas(?) area from application

[SFA]

Hi ,
In the scope of one of our project, we plan to use CEF inside our main application. The application is getting pictures (ie video) from external sensors, process it and then display it. We are searching the most efficient way to directly display the resulting pictures into the rendered page. After reading the API, it is not very clear for us how to do this. Could someone provide hints about that ?

Thanks, SFA

[magreenblatt]

It's not possible to answer your question without more information.

1. Where is the processor located relative to the CEF-based application (e.g. before the CEF-based application, in the application C++ code, or in the application JavaScript code)?

2. If the processor is inside the CEF-based application:
2A. What format does the sensor output image/video in, and at what resolution?
2B. What format does the processor output image/video in, and at what resolution?
2C. What is the available bandwidth between the sensor and the processor?
2D. How does the sensor output image/video data (e.g. local http server, file on disk, in-memory streaming, etc)?
2E. How does the processor output image/video data (e.g. local http server, file on disk, in-memory streaming, etc)?
2F. If the processor will be written in JavaScript, what type of processing are you planning to do?
2G. What is the latency (processing time) of the processor?

3. If the processor is outside the CEF-based application:
3A. What format does the processor output image/video in, and at what resolution?
3B. How does the processor output image/video data (e.g. local http server, file on disk, in-memory streaming, etc)?
3C. What is the available bandwidth between the processor and the CEF-based application?

[SFA]

2. If the processor is inside the CEF-based application:

The video processing occurs in the C++ application, from getting video from external cameras, then performing various pictures enhancement.

2A. What format does the sensor output image/video in, and at what resolution?
2B. What format does the processor output image/video in, and at what resolution?

The resolution is more or less SD (ie 720x576) at 25 fps, but we plan to have HD (1920x1080@50) in the future.

2C. What is the available bandwidth between the sensor and the processor?

This is not an issue, consider infinite bandwidth.

2D. How does the sensor output image/video data (e.g. local http server, file on disk, in-memory streaming, etc)?
2E. How does the processor output image/video data (e.g. local http server, file on disk, in-memory streaming, etc)?

The pictures are directly grabbed/decoded by the C++ application, and then should be displayed immediately, without intermediate stages thru pipes/files/network/...
So in-memory streaming is the right choice, with the minimum number of copies (ie pointer exchange between threads would be the best solution IMHO, if possible)

2F. If the processor will be written in JavaScript, what type of processing are you planning to do?

Nothing occurs on picture at the Javascript level.

2G. What is the latency (processing time) of the processor?

The latency will be 1 frame.

Thanks, SFA

[magreenblatt]

See viewtopic.php?f=10&t=11304 for an overview of your options.