I have used OpenCV to train Haar cascades to detect face and other patterns. However I later realized that Haar tends to give a lot of false positives and I learned of Hog would give a more accurate results. But OpenCV doesn't have a good documentation of how to train hogs, I have googled a bit and found results that includes SVM and others.

OpenCV also has versioning problem where they move certain classes or functions somewhere else.

Are there any other techniques/method that I can use to train and detect objects and patterns? Preferably with proper documentation and basic tutorial/examples. Language preference: C#, Java, C++, Python

  • $\begingroup$ Did you try pyimageseach site? $\endgroup$
    – Ébe Isaac
    Sep 14, 2016 at 6:25
  • $\begingroup$ I did. I was hoping to maybe find some better library. I also read that Hog detection would be slow. So maybe another method or techniques of detecting objects and pattern. $\endgroup$
    – j4rey
    Sep 14, 2016 at 7:04
  • $\begingroup$ Recommending libraries is likely off-topic here, I'm suggesting you to re-ask either at softwarerecs.stackexchange.com or opendata.stackexchange.com $\endgroup$
    – kenorb
    Sep 14, 2016 at 11:10
  • $\begingroup$ thanks......could you recommend a better technique or method for object detection? $\endgroup$
    – j4rey
    Sep 14, 2016 at 12:33

1 Answer 1


The terms you are looking for are deeplearning and convolutional neural networks for object detection. Google responds well to these terms. From academical point of view you can start from:
Single shot multibox detector: https://arxiv.org/pdf/1512.02325v5.pdf
Or Faster-RCNN: https://arxiv.org/pdf/1506.01497.pdf
These are not simple architectures and there are many improvements to them but these give you an idea of the current state of the art methods. There are many implementations of both of these networks in deeplearning libraries for python (eg in Tensorflow, PyTorch).


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .