In this paper we present the Zeitview Rooftop Geometry (ZRG) dataset. ZRG
contains thousands of samples of high resolution orthomosaics of aerial imagery
of residential rooftops with corresponding digital surface models (DSM), 3D
rooftop wireframes, and multiview imagery generated point clouds for the
purpose of residential rooftop geometry and scene understanding. We perform
thorough benchmarks to illustrate the numerous applications unlocked by this
dataset and provide baselines for the tasks of roof outline extraction,
monocular height estimation, and planar roof structure extraction.

本文介绍了 Zeitview Rooftop Geometry（ZRG）数据集，包含成千上万个高分辨率空照图像的采样，用于住宅屋顶几何和场景理解的数字地面模型（DSM），3D 屋顶线框和多视角图像生成点云，并进行了充分的基准测试，以说明此数据集解锁的众多应用程序，并为屋顶轮廓提取、单眼高度估计和平面屋顶结构提取任务提供基线。

ZRG：面向机器学习的高分辨率三维住宅屋顶几何数据集

ZRG: A High Resolution 3D Residential Rooftop Geometry Dataset for  Machine Learning

One significant factor we expect the video representation learning to
capture, especially in contrast with the image representation learning, is the
object motion. However, we found that in the current mainstream video datasets,
some action categories are highly related with the scene where the action
happens, making the model tend to degrade to a solution where only the scene
information is encoded. For example, a trained model may predict a video as
playing football simply because it sees the field, neglecting that the subject
is dancing as a cheerleader on the field. This is against our original
intention towards the video representation learning and may bring scene bias on
different dataset that can not be ignored. In order to tackle this problem, we
propose to decouple the scene and the motion (DSM) with two simple operations,
so that the model attention towards the motion information is better paid.
Specifically, we construct a positive clip and a negative clip for each video.
Compared to the original video, the positive/negative is
motion-untouched/broken but scene-broken/untouched by Spatial Local Disturbance
and Temporal Local Disturbance. Our objective is to pull the positive closer
while pushing the negative farther to the original clip in the latent space. In
this way, the impact of the scene is weakened while the temporal sensitivity of
the network is further enhanced. We conduct experiments on two tasks with
various backbones and different pre-training datasets, and find that our method
surpass the SOTA methods with a remarkable 8.1% and 8.8% improvement towards
action recognition task on the UCF101 and HMDB51 datasets respectively using
the same backbone.

提出了一种解耦场景和物体运动信息的 DSM 方法，通过构造正负剪辑来加强模型对物体运动信息的关注，减少场景信息的影响，并在两项任务上进行实验，发现在 UCF101 和 HMDB51 数据集上动作识别任务的准确率分别提高了 8.1％和 8.8％。