Open anguoyang opened 8 years ago
I tested pedestrian with this command: curl -F source=@"pedestrian.jpg" -F model="pedestrian" localhost:3350/dpm/detect.objects and this is the source image, which saved as pedestrian.jpg: http://special.scol.com.cn/12shijiao/pic/2013/8/5/979452_201385151635_2.jpg with 1200x800 pixels.
It took almost 2 seconds to get the result back.
Conclusion: libccv is only in its very earlier stage for demo purpose, it is far from production, at least for the ccv-serve, yes, it is only a toy.
This toy is easy to learn, the code is so much more readable compare to other non-toy production.
And, it can be used in some speed uncare application, I like it.
I tested pedestrian with this command: curl -F source=@"pedestrian.jpg" -F model="pedestrian" localhost:3350/dpm/detect.objects and this is the source image, which saved as pedestrian.jpg: http://special.scol.com.cn/12shijiao/pic/2013/8/5/979452_201385151635_2.jpg with 1200x800 pixels.
It took almost 2 seconds to get the result back.
Conclusion: libccv is only in its very earlier stage for demo purpose, it is far from production, at least for the ccv-serve, yes, it is only a toy.