Cascade Built In Tracking In Processing

Once I already tested how the face tracking works within my previous post I wanted to let the tracking work for the side part of the face as I feel this is the appropriate path to giving my interactive project to work, this is because when analysing how people walk by in their daily routine within the streets in the public space its always from a side view giving a perfect analysis of data to ensure my design will work for example

People walking

this image shows you the elements of the people moving even though its just a still image of a side shot, i still see this as a prototype of how the tracking could work. By taking this into consideration I wanted the camera to detect the face within this position of the body and face making the image of a hand/eye appear behind them representing the concept of how easy it is to access information from people who are watching them from another location without the person knowing of their existence.

As I already knew about the face tracking of the face I managed to research that the CASCADE has other detecting features for different body parts, this excited me beyond imagination as I was able to find that there was already a reference list of different tracking detection with code already written out, this varied from:

  1. Full Body
  2. Upper Body
  3. Lower Body
  4. Side Face
  5. Left Ear
  6. Right Ear

CASCADE_FRONTALFACE_ALT_TREE

CASCADE_FRONTALFACE_ALT

CASCADE_FRONTALFACE_ALT2

CASCADE_FRONTALFACE_DEFAULT

CASCADE_PROFILEFACE

CASCADE_FULLBODY

CASCADE_LOWERBODY

CASCADE_UPPERBODY

All I had to do was find the one in which was the best for mine in which I found that by changing the FRONTALFACE CASCADE to PROFILEFACE creating the tracking to detect the side part of the face here is the example:

Frontal >>>>PROFILEFACE

This helped me out alot as coding is very difficult for me but as this is open source I already knew that the rest of the code for my project would be my own original due to the fact the PROFILEFACE CASCADE is a built in library from open CV.


 

References

Ubaa.net, n.d. OpenCV::cascade() \ language (API). [online] Available from: http://ubaa.net/shared/processing/opencv/opencv_cascade.html [Accessed 8 Jan. 2015].

Advertisements