采用在线高斯模型的行人检测候选框快速生成方法  

Fast Pedestrian Proposal Generation Algorithm Using Online Gaussian Model

作　　者：覃剑[1] 王美华[1] 

机构地区：[1]重庆大学通信工程学院,重庆400030

出　　处：《光学学报》2016年第11期162-172,共11页Acta Optica Sinica

基　　金：国家自然科学基金(61108086);重庆市自然科学基金(cstc2016shmszx0111);中央高校基金(106112014CDJZR165503)

摘　　要：行人检测是模式识别及机器学习领域的研究热点之一,广泛应用于智能监控、辅助驾驶等领域,而行人候选框的生成是识别及跟踪行人目标的一项重要的前期工作。针对静态监控场景以及特定情况下的车载监控场景,提出了一种基于在线高斯模型的行人检测候选框的快速生成方法(OL_GMPG)。该方法采用高斯模型拟合行人尺寸分布,可以通过生成较少数目的行人候选框达到较高的检测率;并可通过高斯模型的学习与更新过程,获取场景中行人频繁出现的位置以及对应的目标尺度信息,为后续的行人识别及跟踪过程提供辅助。Pedestrian detection is one of the most active research topics in the fields of pattern recognition and machine learning.It has been widely used in intelligent monitoring,auxiliary driving and so on.Generating pedestrian detection proposals is an important work in the early period of pedestrian recognition and pedestrian tracking.Based on the static monitoring scene as well as the on-board monitoring scene under specific circumstances,a novel method to generate pedestrian detection proposals quickly（OL_GMPG）is proposed by using online Gaussian model.High detection rate can be achieved by generating fewer pedestrian detection proposals through the Gaussian model fitting.Both the positions where people appear most frequently and the scale information of corresponding targets can be obtained through the learning and updating processes of the Gaussian model.The information is beneficial to subsequent pedestrian recognition or pedestrian tracking.

关 键 词：机器视觉 行人检测 高斯模型 检测率 尺度信息 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]