There are various popular techniques for the same. The most popular algorithm is by Dalal & Triggs [2005] based on HOG features. They have used SVM for classifying whether a pixel is person or non-person. You can find the article on this link:

lear.inrialpes.fr/people/triggs/pubs/Dalal-cvpr05.pdf

If you are using MATLAB, the newest version of MATLAB has this algorithm pre-implemented as a part of its library support. If you are using OpenCV for your implementation, this algorithm has been coded as the getDefaultPeopleDetector() function.

If your requirements are still not satisfied, I request you to take a look at this:

http://www.mikelrodriguez.com/#/pedestrian-detection-in-videos/

The codes for both these algorithms have been made open-source by their authors.

Dear Vahid Bastani ,

Look the link. May be useful.

chrome-extension://oemmndcbldboiebfnladdacbdfmadadm/http://vision.eecs.ucf.edu/papers/Khan_ECCV06.pdf

Regards, Shafagat

Dear Vahid Bastani,

look the link. May be useful.

chrome-extension://oemmndcbldboiebfnladdacbdfmadadm/https://www.aaai.org/Papers/AAAI/2008/AAAI08-223.pdf

Regards, Shafagat

For detecting the people use HOG and SVM and for tracking purpose, cam shift or mean shift or kalman filter can be used.

Adding to the previous suggestions, for less crowded scenes, check caltech's pedestrian benchmark for assessing the existing algorithms for pedestrian detection in the following links

http://www.vision.caltech.edu/Image_Datasets/CaltechPedestrians/

http://www.vision.caltech.edu/Image_Datasets/CaltechPedestrians/files/algorithms.pdf

http://www.vision.caltech.edu/Image_Datasets/CaltechPedestrians/files/timing.pdf

Apart from the interesting previous comments, you can also follow the ideas presented in our last publication about people tracking in crowded scenes. In this case, our work is focused on video-surveillance in shopping malls. We present a novel robust tracking method for crowded scenes:

http://dx.doi.org/10.1016/j.eswa.2015.06.016

https://www.youtube.com/watch?v=Rqk-vFKzAcQ

Good luck with your research!

Article Expert Video-Surveillance System for Real-Time Detection of ...

Mr. Bastani

I've been working with Weightless Neural Networks for both Detection and Tracking of scene objects. 

A full paper on Visual Tracking with WNN is on the verge of beign published, however there are already available papers regarding the usage of WNN for Traffic Sign detection and recognition. Most of the concepts for the WNN-based Visual Tracking were incorporated from these papers.

Hope it helps you on your research

Best Regards.

Conference Paper Traffic sign detection with VG-RAM Weightless Neural Networks

Conference Paper Traffic sign recognition with VG-RAM Weightless Neural Networks