Open AccessLandmark Points Detection in Case of Human Facial Tracking and Detection
Author(s) -
Anita Jindal,
Rashmi Priya
Publication year - 2019
Publication title -
international journal of engineering and advanced technology
Language(s) - English
Resource type - Journals
ISSN - 2249-8958
DOI - 10.35940/ijeat.b3367.129219
Subject(s) - landmark , artificial intelligence , computer vision , computer science , face detection , support vector machine , facial expression , three dimensional face recognition , face (sociological concept) , pattern recognition (psychology) , haar like features , histogram , biometrics , chin , facial recognition system , classifier (uml) , facial motion capture , histogram of oriented gradients , object class detection , image (mathematics) , medicine , social science , sociology , anatomy
This paper describes the human facial landmark points detection is very important in the field of image processing as face detect, face identifies, face re-construct, face corners alignment, different head pose and facial expression analysis. Facial landmark is an essential point for applying face processing operation ranging from biometric recognition to mental states. In this paper, Haar cascading face detection technique is used to face detection and tracking. Histogram of Oriented Gradients (hog) has been used for 68 landmark points detection in case of human tracking and detection and support vector machine (svm) classifier are used for 68 landmark points detection for right-left eyebrow, left-right eye, nose, lips, chin, and jaw. The existing methods work effectively but many issues occur in detection as of different head poses, facial expressions, facial occlusion, illumination, colour, shadowing and self-shadowing etc. The performance of experimental results shows the advantages of our purposed method is highly accurate in terms of facial 68 landmark points tracking and detection and less error detection rate with the Multi-PIE database.