公开(公告)号：US20170032676A1

公开(公告)日：2017-02-02

申请号：US15216934

申请日：2016-07-22

Applicant: Maral MESMAKHOSROSHAHI ,  Maziar LOGHMAN ,  Joohee KIM

Inventor： Maral MESMAKHOSROSHAHI ,  Maziar LOGHMAN ,  Joohee KIM

IPC: G08G1/16 ,  G06T7/00 ,  H04N13/02

CPC classification number: H04N13/204 ,  G06K9/00369 ,  G06K9/00805 ,  G06K9/3233 ,  G06K9/6289 ,  G06T7/11 ,  G06T2207/10012 ,  G06T2207/10028 ,  G06T2207/30261 ,  H04N13/239 ,  H04N2013/0081

Abstract: A region of interest (ROI) generation method for stereo-based pedestrian detection systems. A vertical gradient of a clustered depth map is used to find ground plane and variable-sized bounding boxes are extracted on a boundary of the ground plane as ROIs. The ROIs are then classified into pedestrian and non-pedestrian classes. Simulation results show the algorithm outperforms the existing monocular and stereo-based methods.

Abstract translation: 用于立体声行人检测系统的感兴趣区域（ROI）生成方法。 使用聚类深度图的垂直梯度来查找地平面，并且在地平面的边界上提取可变大小的边界框作为ROI。 然后将ROI分类为行人和非行人班。 仿真结果表明，该算法优于现有的单目和立体声方法。

公开(公告)号：US09754377B2

公开(公告)日：2017-09-05

申请号：US14827897

申请日：2015-08-17

Applicant: Maziar Loghman ,  Maral Mesmakhosroshahi ,  Joohee Kim

Inventor： Maziar Loghman ,  Maral Mesmakhosroshahi ,  Joohee Kim

IPC: G06K9/00 ,  G06T7/00 ,  G06T5/00 ,  G06T7/593 ,  H04N13/00

CPC classification number: G06T7/593 ,  G06T5/005 ,  G06T2207/10012 ,  G06T2207/10028 ,  G06T2207/20028 ,  G06T2207/20192 ,  H04N2013/0081

Abstract: A computer-implemented depth estimation method based on non-parametric Census transform with adaptive window patterns and semi-global optimization. A modified cross-based cost aggregation technique adaptively creates the shape of the cross for each pixel distinctly. In addition, a depth refinement algorithm fills holes within the estimated depth map using the surrounding background depth pixels and sharpens the object boundaries by exerting a trilateral filter to the generated depth map. The trilateral filter uses the curvature of pixels as well as texture and depth information to sharpen the edges.

公开(公告)号：US20160048970A1

公开(公告)日：2016-02-18

申请号：US14827897

申请日：2015-08-17

Applicant: Maziar Loghman ,  Maral Mesmakhosroshahi ,  Joohee Kim

Inventor： Maziar Loghman ,  Maral Mesmakhosroshahi ,  Joohee Kim

IPC: G06T7/00 ,  H04N13/02 ,  G06T5/00

CPC classification number: G06T7/593 ,  G06T5/005 ,  G06T2207/10012 ,  G06T2207/10028 ,  G06T2207/20028 ,  G06T2207/20192 ,  H04N2013/0081

Abstract: A computer-implemented depth estimation method based on non-parametric Census transform with adaptive window patterns and semi-global optimization. A modified cross-based cost aggregation technique adaptively creates the shape of the cross for each pixel distinctly. In addition, a depth refinement algorithm fills holes within the estimated depth map using the surrounding background depth pixels and sharpens the object boundaries by exerting a trilateral filter to the generated depth map. The trilateral filter uses the curvature of pixels as well as texture and depth information to sharpen the edges.

Abstract translation: 基于非参数普查变换的计算机实现的深度估计方法，具有自适应窗口模式和半全局优化。 改进的基于交叉的成本聚合技术自动地为每个像素明确地创建十字的形状。 此外，深度细化算法使用周围的背景深度像素在估计的深度图内填充空洞，并通过向生成的深度图施加三边形过滤器来锐化对象边界。 三边滤镜使用像素的曲率以及纹理和深度信息来锐化边缘。