人工智能培训

搜索

深度学习论文:板球中风提取:建立大型板球动作数据集(Cricket stroke extraction: Towards creation of a large

[复制链接]
755144030 发表于 2019-1-11 11:08:42 | 显示全部楼层 |阅读模式
755144030 2019-1-11 11:08:42 189 0 显示全部楼层
深度学习论文:板球中风提取:建立大型板球动作数据集(Cricket stroke extraction: Towards creation of a large-scale cricket  actions dataset)在本文中,我们处理大规模未修剪的板球视频数据集的时间动作定位问题。我们对板球视频感兴趣的动作是由击球手进行的板球击球,通常由放置在板球场两端的板球场看台上的相机覆盖。应用一系列预处理步骤后,我们在数据集中以恒定的帧速率和分辨率为1110个视频提供了大约7300万帧。定位方法是一种广义的方法,它应用CUT检测的带有随机森林模型(使用总和灰度组合图差异特征)和两个线性SVM相机模型(CAM1和CAM2)用于第一帧检测,训练CAM1和CAM2视频的HOG特征。假设CAM1和CAM2是板球击球的一部分。在预测的边界位置处,计算第一帧的HOG特征,并且使用简单的算法来组合正向预测的摄像机。为了使过程尽可能通用,我们没有考虑领域特定的知识,例如跟踪或特定的形状和运动特征。我们的方法的详细分析与用于评估单个模型的指标一起提供,并且最终预测段。在测试集的小样本上实现了加权平均TIoU为0.5097。
In this paper, we deal with the problem of temporal action localization for alarge-scale untrimmed cricket videos dataset.Our action of interest forcricket videos is a cricket stroke played by a batsman, which is, usually,covered by cameras placed at the stands of the cricket ground at both ends ofthe cricket pitch.After applying a sequence of preprocessing steps, we have~73 million frames for 1110 videos in the dataset at constant frame rate andresolution.The method of localization is a generalized one which applies atrained random forest model for CUTs detection(using summed up grayscalehistogram difference features) and two linear SVM camera models(CAM1 and CAM2)for first frame detection, trained on HOG features of CAM1 and CAM2 videoshots.CAM1 and CAM2 are assumed to be part of the cricket stroke.At thepredicted boundary positions, the HOG features of the first frames are computedand a simple algorithm was used to combine the positively predicted camerashots.In order to make the process as generic as possible, we did not considerany domain specific knowledge, such as tracking or specific shape and motionfeatures.The detailed analysis of our methodology is provided along with the metricsused for evaluation of individual models, and the final predictedsegments.Weachieved a weighted mean TIoU of 0.5097 over a small sample of the test set.深度学习论文:板球中风提取:建立大型板球动作数据集(Cricket stroke extraction: Towards creation of a large-scale cricket  actions dataset) NwmIoUvvZMzhK1bm.jpg
URL地址:https://arxiv.org/abs/1901.03107     ----pdf下载地址:https://arxiv.org/pdf/1901.03107    ----深度学习论文:板球中风提取:建立大型板球动作数据集(Cricket stroke extraction: Towards creation of a large-scale cricket  actions dataset)
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则 返回列表 发新帖

755144030当前离线
新手上路

查看:189 | 回复:0

快速回复 返回顶部 返回列表