Is Haar Cascade the only image recognition method available in OpenCV

Question

I know that in OpenCV there are many detection methods, such as SURF, STAR, ORB, etc., but these methods are designed to detect the function of a new video stream, and not to access specific instances of objects that require prior training. The OpenCV documentation is not so easy to turn over, and I have not yet been able to find anything other than Haar, which, as I know, is best suited for face recognition.

Are there any other methods besides Haar? The Haar technique dates back to research 10 years ago, so ideally I hope that since then the additional successes that have been implemented in OpenCV have been achieved.

Answer 1

If you are looking for algorithms such as OpenCV machine learning, check out this link .

For a modern on-the-fly object detection algorithm, take a look at OpenTLD. It uses bounding fields and random forests to learn about the object over time. Check out the demo video here .

Also check out the matching_to_many_images.cpp sample from OpenCV. It uses function descriptors to map objects similar to Google Goggles. An example of this is the bagofwords_classification.cpp example. This may be what you are looking for in this case. It uses function detectors (SURF, SIFT, etc.) to detect objects, and then classifies them by comparing the relative positions of these functions with a recognizable database of functions. See also the tutorial from MIT.

Latentsvmdetect.cpp could also be a good starting point for you.

Hope this helps!