Pain Detection in Children using Image Processing

Abstract: Pain is a qualitative character which is very difficult to assess and manage. In younger children it is not possible to measure pain by self report. Behavioral scientists say that pain can be measured by using the facial indicators. The existing approach for assessing pain in children requires the services of a skilled staff. The output thus obtained may not be accurate enough and has a subjective bias also. Here we have suggested a method for assessing the pain in children using image processing techniques by considering facial indicators. Using Viola Jones algorithm the image of the face is segmented and the features are extracted. The classification is done with the help of multi SVM classifier. We obtained a reliability of 0.97 for our system by interrator method, which is very high. This technique is very useful for assessing the pain in children in post operative wards, Intensive Care Units etc.

Keywords: Pain, children, pain assessment, image processing, Viola Jones, Multi SVM.