Human Detection for a Video Surveillance Applied in the ‘SmartMonitor’ System

摘要: Human detection is one of the key and crucial tasks in video surveillance systems important for purpose object tracking, fall detection, human gait analysis or abnormal event detection. This paper concerns application two classifiers ‘SmartMonitor’ system — an intelligent security based on image analysis. The are Histogram Oriented Gradients (HOG) descriptor simple Haar-like features. provides a brief description main characteristics, discusses problem includes some results experiments performed using various parameters HOG Haar that were trained benchmark databases tested appropriate sequences. aims at investigating effectiveness performance both methods applied separately before incorporating them into system’s processing model.