全景视频的目标跟踪方法、可读存储介质及计算机设备

    公开(公告)号:WO2021139787A2

    公开(公告)日:2021-07-15

    申请号:PCT/CN2021/070922

    申请日:2021-01-08

    Inventor: 许睿 姜文杰

    Abstract: 本申请适用于视频处理领域,提供了一种全景视频的目标跟踪方法、可读存储介质及计算机设备。所述方法包括:采用所述跟踪器对所述待跟踪目标进行跟踪检测,得到所述待跟踪目标在下一全景视频帧的预测跟踪位置,计算所述预测跟踪位置的置信度,并采用遮挡检测器计算所述预测跟踪位置的遮挡分数;判断所述预测跟踪位置的置信度是否大于预设置信度阈值,并判断所述预测跟踪位置的遮挡分数是否大于预设遮挡分数阈值;根据置信度和遮挡分数采取相应的跟踪策略。本申请能够区分跟踪失败的原因是由于目标丢失还是遮挡导致,进而采取相应的跟踪恢复策略,在跟踪失效时能自动恢复跟踪,从而达到长时间持续跟踪的效果,且本发明方法具有较低的运算复杂度,实时性好。

    FACIAL SYNTHESIS FOR HEAD TURNS IN AUGMENTED REALITY CONTENT

    公开(公告)号:WO2022213099A1

    公开(公告)日:2022-10-06

    申请号:PCT/US2022/071456

    申请日:2022-03-31

    Abstract: The subject technology receives frames of a source media content, the frames of the source media content including representations of a head and a face of a source actor. The subject technology generates, based at least in part on the frames of the source media content, sets of source pose parameters. The subject technology receives at least one target image, the at least one target image including representations of a target head and a target face of a target entity. The subject technology provides the sets of source pose parameters to a neural network to determine facial landmarks for head turns and facial expressions. The subject technology generates, based at least in part on the sets of source pose parameters and the facial landmarks for head turns and facial expressions, an output media content. The subject technology provides augmented reality content based at least in part on the output media content for display on a computing device.

    一种改进的倾斜矩形范围框标注方式

    公开(公告)号:WO2022007943A1

    公开(公告)日:2022-01-13

    申请号:PCT/CN2021/105454

    申请日:2021-07-09

    Abstract: 本发明公开了一种改进的倾斜矩形范围框标注方式,标注方式中用于标注的量是:中心点C的坐标、中心点到任意一个顶点D的向量(I)、C到D的一个相邻顶点E的向量(II)在(I)上的投影向量(III)与(I)的比例系数;要求(III)与(I)同向以及从(I)到(II)的夹角只能是顺时针方向或者逆时针方向中的一种;另规定记录在第一个位置的标注向量(I)的分量到该向量的夹角是顺时针或逆时针方向且这个夹角取值范围为[0,90),第二个位置可以记录标注向量的另一个分量或者标注向量的模,在第三个位置上记录第一个分量的方向,该方向可以取X轴方向或Y轴方向,范围框是正方形时既可以取X轴方向也可以取Y轴方向。该标注方式为正方形和一般矩形范围框采用了完全相同的外部约束,有利于机器学习算法识别外部约束。

    TARGET RE-IDENTIFICATION METHOD, NETWORK TRAINING METHOD THEREOF, AND RELATED DEVICE

    公开(公告)号:WO2022001034A1

    公开(公告)日:2022-01-06

    申请号:PCT/CN2020/139349

    申请日:2020-12-25

    Abstract: A target re-identification method, a network training method thereof and a related device. The training method comprises the steps of obtaining a training image set; identifying each training image in the training image set by using a target re-identification network; obtaining an identification result of each training image; wherein the target re-identification network comprises a plurality of branches; wherein the recognition result of each training image comprises feature information output by each branch and a classification result corresponding to the feature information; wherein the feature information output by one branch comprises n pieces of local feature information, n is greater than 3, and the n pieces of local feature information correspond to different image areas of the training image; and adjusting the parameter of each branch of the target re-identification network based on the identification result of the training image. In this way, the result of target recognition by the trained target re-recognition network is more accurate.

    基于场景识别的车辆自适应传感器系统

    公开(公告)号:WO2021253741A1

    公开(公告)日:2021-12-23

    申请号:PCT/CN2020/133599

    申请日:2020-12-03

    Abstract: 一种基于场景识别的车辆自适应传感器系统,包括:AVM环视传感器(1)、AVM环视控制器(3)及ADAS控制器(5);AVM环视传感器(1)包含摄像头、位置调整机构(11)和位置传感器(12);摄像头向AVM环视控制器(3)传递采集到的图像信息,位置调整机构(11)接收AVM环视控制器(3)的位置调整指令并调整摄像头的位置,位置传感器将摄像头的位置信息向AVM环视控制器(3)反馈;AVM环视控制器(3)向ADAS控制器(5)传递识别到的目标信息。上述系统通过场景识别优化,得到适应于当前行车场景的ROI感兴趣区域,提高系统的图像处理性能;在AVM环视传感器(1)中对图像进行目标识别,输出的目标和场景信息可用于ADAS控制器(5),能降低ADAS控制器(5)所需的图像处理芯片能力,同时能够降低ADAS系统的功耗,有效地降低配置成本。

Patent Agency Ranking