We are getting there, with as usual some trade-off between quality and speed.

For example [Lecture 8: Spatial Localization and Detection lecture][1] shows some benchmarks (mAP = Mean Average Precision, higher is better; FPS = frame per second):

[![Table/Performance: Real-Time Detectors][2]][2]

[![Table/Performance: R-CNN, Fast R-CNN, Faster R-CNN][3]][3]


  [1]: http://cs231n.stanford.edu/slides/2016/winter1516_lecture8.pdf
  [2]: https://i.sstatic.net/AfHt2m.png
  [3]: https://i.sstatic.net/B5Qb9m.png