High Impact Factor : 4.396 icon | Submit Manuscript Online icon |

An Intelligent way of searching crime scenes on a video surveillance system


Mr.B.Naresh Kumar , AVS Engineering College, Salem-636003; Mr.I.Selvamani, AVS Engineering College, Salem-636003.


Feed Forward Neural Network, Principal Component Analysis, Video Surveillance System.


With the increasing demands of visual surveillance systems, vehicle and people identification at a distance have gained more attention for the researchers recently. Extraction of information from images and image sequences are very important for the analysis according to the application. We present an image processing software suite, based on the Mat lab environment, specifically designed to be used as a forensic tool by law enforcement laboratories in the analysis of crime scene videos and images. Here the normal and abnormal actions of a human in any of the under surveillance place is determined and its edges been traced out with various image segmentation techniques. From the preprocessed image database feature vectors are extracted using two methods namely receptive field method and features from Gabor Filter. Features extracted from the database is used as an input of the classifier, which is a feed forward neural network (shortly called FFNN) on a reduced subspace learned by an approach simpler than principal component analysis (shortly called PCA). Once the neural network has been trained, the database will be classified into two categories: normal faces and suspicious faces. Thus suspicious frames are extracted which in turn make ease the process of crime identification.

Other Details

Paper ID: IJSRDV2I4057
Published in: Volume : 2, Issue : 4
Publication Date: 01/07/2014
Page(s): 97-101

Article Preview

Download Article