| |
 |
UNIVERSITY OF BUCHAREST
FACULTY OF PHYSICS Guest
2024-11-23 18:25 |
 |
|
|
|
Conference: Bucharest University Faculty of Physics 2017 Meeting
Section: Biophysics; Medical Physics
Title:
Detection and recognition of human patterns
Authors:
Theodora PREDA, Radu MUTIHAC
Affiliation:
University of Bucharest, Faculty of Physics, PO Box MG-38, Bucharest-Magurele, Romania
E-mail
theodoradanielapreda@gmail.com
Keywords:
Face detection, pattern recognition, image analysis, machine learning
Abstract:
These days, almost any owner of a gadget with a camera, observes the automatic appearance of a rectangle or circle drawing that outline of the human faces towards which the objective of the camera is pointed to. This research work describes a pattern detection and recognition approach based on Viola-Jones algorithm, which is capable of processing images very quickly while reaching high detection rates. From the various patterns modeling techniques, the statistical approach has been thoroughly studied and used most often in practice. More recently, the addition of a theory of artificial neural techniques has received more attention. The design of a recognition system requires special attention to the following: definition of the model class, model representation, extraction of features, classification learning, training selection, and the evaluation of performance. The main objectives of our work have been to present the issue of automatic face detection with a brief overview of the specific difficulties and main types of approaches to solving the detection and recognition tasks. Likewise, we have been searching to detail at a level that aims to be as accessible as possible one of the most known and widely used method, namely the Viola-Jones algorithm, which is the first face detection system in real time. The importance of the selected subject resides in a multitude of potential applications in science and our society, specifically in monitoring airports, railway stations, metro stations, supermarkets, generally the environments where several persons are an easy target for terrorist mass attacks.
|
|
|
|