1. List
Learning where to attend like a human driver: Intelligent Vehicles Symposium (IV), 2017 IEEE
1.1 2D CNN
1.2 Point Cloud
2. Paper
[Computer Vision for Autonomous Vehicles: Problems, Datasets and State-of-the-Art]- Deep Semantic Segmentation for Automated Driving: Taxonomy, Roadmap and Challenges
2.1 2D CNN
MultiNet: Real-time Joint Semantic Reasoning for Autonomous Driving: classifi- cation, detection and semantic segmentation Task수행시 값을 공유 하여
속도 향상에 초점, 2016Deep convolutional neural networks for pedestrian detection: 보행자 탐지, 2016
Unified multi-scale CNN. (KITTI: 8th car, 1st ped) [Cai, ECCV '16] [Home] [Code] [Video]
Subcategory-aware CNN. (KITTI: 7th car, 3rd ped)) [Xiang, Arxiv '16] [Home]
- Exploit all layers. (KITTI: 10th car, 5th ped) [Yang, CVPR '16] [Home]
- Visual Object Recognition with 3D-Aware Features in KITTI Urban Scenes [Yebes, Sensors '15]: stereo-vision
2.2 Multiview
[A Comparative Analysis and Study of Multiview CNN Models for Joint Object Categorization and Pose Estimation](http://proceedings.mlr.press/v48/elhoseiny16.pdf): 추가자료, 3D 데이터 관련 내용인지 재 확인 필요
Vehicle Detection from 3D Lidar Using Fully Convolutional Network ~~[[Li, RSS '16]~~](http://www.roboticsproceedings.org/rss12/p42.pdf)
2.3 Point Cloud
Object Classification using 3D Convolutional Neural Networks: 2016
OctNet: Learning Deep 3D Representations at High Resolutions: 2016~2017
Deep Semantic Classification for 3D LiDAR Data: 물체를 고정된, 움직이는, 움직일수 있는 것으로 3분류
Generalized Convolutional Neural Networks for Point Cloud Data: 2017
Fast LIDAR-based Road Detection Using Fully Convolutional Neural Networks: 2017.03
[sparse convolutional neural networks]
Alexandr Notchenko, Ermek Kapushev, Evgeny Burnaev. Sparse 3D Convolutional Neural Networks for Large-Scale Shape Retrieval. CoRR, 2016.
M. Engelcke, D. Rao, D. Zeng Wang, C. H. Tong, I. Posner. Vote3Deep: Fast Object Detection in 3D Point Clouds Using Efficient Convolutional Neural Networks. CoRR, 2016.
Ben Graham. Sparse 3D convolutional neural networks. BMVC, 2015.
CNN-Based Object Segmentation in Urban LIDAR with Missing Points: 2016
LIDAR point upsampling. [Schneider, Arxiv '16]
2.4 Fusion
- 2D/3D Sensor Exploitation and Fusion for Enhanced Object Detection (Similar to ours) [Xu, CVPRW '14]
3. Article (Post, blog, etc.)
3. Tutorial (Series, )
4. Youtube
- Tutorial : 3D Deep Learning: CVPR 2017, MVCNN, 3DCNN, Point Cloud, Meshes
6. Material (Pdf, ppt)
ADAS (Advanced Driving Assistance System): ppt 261, Baidu USA, 2017.04
6.2 Point Cloud
Synthesize for Learning: Joint analysis of 2D images and 3D shapes: ShapeNet (3D Datasets)3D Deep Learning on Geometric Forms: [추천] 3D 모델들에 대한 설명, 3D DL 설명
Scene Understanding with 3D Deep Networks: 3D Match, Deep Sliding Shapes, SSCNet(Semantic Scene Completion)Learning 3D representations,disparity estimation, and structure from motion : FlowNet, DispNet, DeMoN
AutonomousDrivingCookbook: MS, 자율주행 자동차와 관련한 AirSim 시뮬레이션 환경과 Keras 를 이용해 관련 프로젝트를 학습하고 프로젝트를 수행
7. Implementation (Project)
- PyDriver: Training and evaluating object detectors and classifiers in road traffic
Project: Development of AEB System for Pedestrian Protection (4th year)
Kaggle_3D MNIST: A 3D version of the MNIST database of handwritten digits
7.1 시뮬레이션 툴
CARLA is an open-source simulator for autonomous driving research. 설명
Reading game frames in Python with OpenCV - Python Plays GTA V: GTA를 이용한 자율 주행
8. Research Group / Conference
DeepDrive: 버클리, 3D Object Detection based on Lidar and Camera Fusion
Improving 3D Perception for Object Detection, Classification and Localization using Fused Multimodal Sensors: Michigan State University
- Researcher : Saif Imran, Aaron Gonzalez, Mehmet Akif Alper
IEEE Intelligent Vehicles Symposium : 2017