1) A system for the automatic visual estimation of in-class polling using an iPhone is proposed
which detects the number of faces within the field of view,
2) The number of raised hands within the field of view, and uses the ratio as the estimate of the
number of positive votes.
3) The system utilizes the Adaboost object detection algorithm implemented in OpenCV, and
processes all detected faces and raised hands, to determine whether each of the audience
available in the scene are voting or not.
4) The system works as a standalone iPhone application. So, the whole process of audience
polling is handled by a single iPhone.
Algorithms And Framework
1) Adaboost Algorithm : It use for face detection and compute the face location.
2) OpenCV(Open Source Compute Vision) : A library of programming functions mainly aimed at
real-time computer vision.it use open framework.
3) ROI(Region Of Interest) : A selected subset of samples within a dataset identified for a
Fig. 1.Sample images of an audience (photos are taken with iPhone).
1) The regions of the image that pass all layers of this cascade classifier
are then considered as detected objects.
Fig .2.Sample images of an audience (photos are taken with iPhone).
1) Before starting the main
procedure for finding raised
hands in the image, the image
is processed to find the faces.
2) Its Detect the face through
Fig. 3. Faces detected using Adaboost algorithm(N==8)
3) Region of interest (ROI) for finding hands based on the detected
Fig. 4. Faces with Hand detected using
Finding Raised Hands
1) OpenCV has the ability to train
a new hand detector if it is
provided with a large number
(1,000–10,000) of positive
(hand) and negative images.
Fig.5. Aligning all hand images with a reference image through
a rigid-body spatial registration.
1) Used OpenCV (Cross Platform ):Using Objective-C Language
2) Facial recognition system.
3) Gesture recognition.
4) Image processing.
1) implemented a completely automatic system for polling an audience using
iPhone. The whole process of audience polling is handled by a single iPhone.
2) The system then finds the people and their hands in the image and reports
the polling results.
3) The complete audience polling system has been successfully implemented on
iPhone where the whole process of polling the audience takes Formula
seconds to complete
1) “Automated audience polling on iPhone" Electrical and Computer
Engineering (CCECE), 2013 26th Annual IEEE Canadian Conference on, Issue
Date: 5-8 May 2013, Written by: Heidari, A.; Aarabi, P.
2) Paul Viola and Michael J. Jones, “Robust real-time face detection”
International Journal of Computer Vision, 2004.
3) H. Nolker, C.; Ritter, “Visual recognition of continuous hand postures” Neural
Networks, IEEE Transactions on,2002.