三维场景感知之三维目标检测方向入门

1 文档需知

本文档为实验室整理，仅供学习使用。未经允许严禁私自转载。

同时欢迎大家报名加入实验室：厦门大学ASC实验室

三维目标检测入门

2 基础知识

• 机器学习（机器学习（西瓜书）周志华， 吴恩达系列视频）
• 深度学习 (李飞飞系列视频，各种深度学习实战书籍)
• Python 编程基础 (网易云课堂搜索 python 基础)
• Pytorch 深度学习框架 (pytorch 实战书籍)

深度学习基础必上手项目

• MNIST 手写数字识别
• Cifar10 图像分类项目

科研研究必知道的论文门户

• 谷歌学术：https://scholar.google.com
• Arxiv：https://arxiv.org/
• 语义学术：https://www.semanticscholar.org/
• 微软学术：https://academic.research.microsoft.com/

深度学习必看论文

• VGG：Very Deep Convolutional Networks for Large-Scale Image Recognition
• ResNet： Deep Residual Learning for Image Recognition
• BatchNorm：Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Adam：Adam: A Method for Stochastic Optimization
• LSTM：Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting
• Transformer：Attention Is All You Need

3 目标检测入门知识

二维目标检测必看论文

• R-CNN
• Fast R-CNN
• Faster R-CNN
• YOLO 系列
• SSD 系列

4 三维目标检测入门知识

三维目标检测必熟悉数据集

• KITTI
• Nuscence
• Waymo

三维目标检测点云分类分割预备知识

• PointNet (必手撕代码)
• PointNet++

三维目标检测必熟悉，必跑通，必理解细节实现的框架

• OpenPCDet
• Det3D

三维目标检测必看（有时间必复现）论文（都有开源代码）

• SECOND：Sparsely Embedded Convolutional Detection
• VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection
• PointRCNN: 3D Object Proposal Generation and Detection from Point Cloud
• PointPillars: Fast Encoders for Object Detection from Point Clouds
• 3DSSD: Point-based 3D Single Stage Object Detector
• SA-SSD：Structure Aware Single-stage 3D Object Detection from Point Cloud
• PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection
• Voxel R-CNN: Towards High Performance Voxel-based 3D Object Detection
• Center-based 3D Object Detection and Tracking
• SE-SSD: Self-Ensembling Single-Stage Object Detector From Point Cloud
• Voxel Transformer for 3D Object Detection
• Improving 3D Object Detection with Channel-wise Transformer
• Behind the Curtain: Learning Occluded Shapes for 3D Object Detection
• Sparse Fuse Dense: Towards High Quality 3D Detection with Depth
  Completion
• DSVT: Dynamic Sparse Voxel Transformer with Rotated Sets
• SAFDNet: A Simple and Effective Network for Fully Sparse 3D Object Detection