VIDEO IMAGE DETECTION AND TRACKING USING PCA AND CONTOUR MAPPING SCHEME WITH HARDWARE IMPLEMENTATIONAND DESIGN

Abstract

newlineABSTRACT newline newline newline newline newlineFace detection and recognitions has an important role in cyber and surveillance and patient monitoring system and implementation constraints limit their performance As compared with other biometrics systems like fingerprint palm print and iris based recognition this work has distinct advantages due to its non invasive and non contact process Face images are captured from a distance and the identification does not require interaction with the subject Similarly sliced video inputs has received much attention due to its ability to preserve data privacy and still support indexing searching mining and other required operations essential to medical domain In this context in this research work novel algorithms interleaved with h sliced algorithm are presented with focus on faster retrieval schemes to suit realtime implementation This proposed work is applied to patient monitoring systems to detect backward and forward fall of patients in hospitals Pose variation creates an artifact in face detection and recognition This effect is taken care to detect the images and recognizes the correct image even with changes in the posture newlineNevertheless existing fall detection research is facing various limitations This study aimed to develop and validate a new fall detection algorithm using 2D information ie trunk angular velocity and trunk angle The angular kinematics was measured using inertial measurement unit during slip induced backward falls and a variety of daily activities The new algorithm was on average able to detect backward falls prior to impact with ninety five percentage sensitivity ninety percentage specificity and two fifty five ms response time Hence it is concluded that the fall detection algorithm detect falls during motion and is suited for the elderly population newlineTo realize the fall detection algorithm, the angles acquired by the sensors are used the system inputs and a realtime system on Linux OS is developed to process and make a detection quickly OpenCVbased API,s are used effectively and the work can be easily ported into a handheld embedded device To make the system scalable and ported with minimum memory requirement dimensional reduction algorithm PCA based intrinsically Coexist The realtime system generates an ordered feature sequence and then examined in a sequential manner by the proposed nonlinear classifier for recognition purpose This process iterates for every frame of feature sequence newline newlineKeywords Fall detection Realtime systems Classifier open CV based MC vision newlineFace recognition PCA algorithm newline newline