Shinning

A slow learner's Blog

Sce Loss

Symmetric Cross Entropy Learning (SL) For Segmentation Code for ICCV2019 “Symmetric Cross Entropy for Robust Learning with Noisy Labels” https://ar...

Posted by Packy on July 30, 2020

Binary Search

二分搜索一、寻找一个数（基本的二分搜索）** 这个场景是最简单的，肯能也是大家最熟悉的，即搜索一个数，如果存在，返回其索引，否则返回 -1。 int binarySearch(int[] nums, int target) { int left = 0; int righ...

Posted by Packy on June 1, 2020

Si3d

S3DIS Our dataset is composed of five large-scale indoor areas from three different buildings, each covering approximately 1900, 450, 1700, 870 an...

Posted by Packy on May 20, 2020

Super Point Graph

Superpoint-graph的思想，对点云进行预先的分割，分割为几何类似的面片。同时利用RGCN 少量监督信息，对大规模点云进行语义分割。 Super-point graph: Large-scale Point Cloud Semantic Segmentation with ...

Posted by Packy on May 17, 2020

Pointcloud Interactive Seg

点云交互式分割调研 1 Golovinskiy, A., & Funkhouser, T. A. (2009). Min-cut based segmentation of point clouds. ICCV Workshops, 39–46. http://doi.org/10....

Posted by Packy on May 15, 2020

2018_nsr_a brief introduction to weakly supervised learning

摘要：监督学习技术通过学习大量训练样本来构建预测模型，其中每个训练样本都有一个标签标明其真值输出。尽管当前的技术已经取得了巨大的成功，但是值得注意的是，由于数据标注过程的高成本，很多任务很难获得如全部真值标签这样的强监督信息。因此，能够使用弱监督的机器学习技术是可取的。本文综述了弱监督学习的一...

Posted by Packy on May 12, 2020

Iccv19reading

ICCV 2019 文献读后感找能解决问题的文献。直接处理点云的网络（除开体素化、多视图）基本分为两派：延续pointnet的变体： Pointnet++、 shellNet、 DGCNN、 geometry sharing net 等等基于稀疏卷积的： 2019_I...

Posted by Packy on March 15, 2020

H5 Pointcloud Convert To Pyg Dataset

#将h5格式点云数据集转化为PyG DataSet 导入点云数据集 def load_data(partition): download() BASE_DIR = os.path.dirname(os.path.abspath(__file__)) DATA_DIR = ...

Posted by Packy on March 14, 2020