c++ - Best approach to detect people in a STATIC camera images ( not using video stream ) - HOG, Haar, BackgroundSubtract in OpenCV? -
So I'm planning to implement static camera images to human identities - the camera never runs, but light The situation of the person may be right or sitting.
Question:
In such a case, what would be better for people to find out?
Horsescade is mostly related to the face and eyes and my images are taken from an angle, especially for those on the back side of the camera It is very difficult to trace the characteristics.
HOG Features - I'm sure this one
backgroundSubtractorMOG - does not seem to provide a very good result now Even by using several images to "create a background, to run it through subteacacters"
Sample image with no one in it:
It would be easy to find out that a picture And runs? Detecting the body, which can be in different positions, will be very difficult (especially if half of the body is hidden behind a table) I am writing a face detection algorithm in the moment using Horsscade - this is someone It can also deal with the type of size (not only the face and eyes), but some general characteristics should be found between those items - the body can vary greatly in colors, shapes. It will also be difficult to create a waterfall (you will need thousands of sample pictures)
If the camera is static, then things going forward will be human use that fact I have no experience in this matter, but This place can signal to you:
Comments
Post a Comment