Object-Detection-Development

Outline

1. Classical papers
2. Benchmarks
3. Great open code
4. Extra Reading notes
5. Very Recent Works

1. Classical papers

• 2013 RCNN --Rich feature hierarchies for accurate object detection and semantic segmentation

• 2013 ----Deep Neural Networks for Object Detection

• 2014 SPPnet --Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition-sppnet

• 2015 Cascaded Sparse Spatial Bins for Efficient and Effective Generic Object Detection

• 2015 Fast R-CNN

• 2016 Faster R-CNN --Towards Real-Time Object Detection with Region Proposal Networks

• 2016 Inside-Outside Net_Detecting Objects in Context with Skip Pooling and Recurrent Neural Networks

• 2016 R-FCN --Object Detection via Region-based Fully Convolutional Networks

• 2016 SSD --SSD: Single-Shot MultiBox Detector

• 2016 PVANet --PVANET: Deep but Lightweight Neural Networks for Real-time Object Detection

• 2016 YOLO9000 --YOLO9000: better,faster,stronger

• 2016 YOLO --You Only Look Once-Unified, Real-Time Object Detection

• 2017 A-Fast-RCNN --A-Fast-RCNN_Hard Positive Generation via Adversary for Object Detection

• 2017 Speed_Accuracy trade-offs for modern convolutional object detectors

• 2017 DSSD --DSSD : Deconvolutional Single Shot Detector

• 2017 DeformableCNN --Deformable Convolutional Networks

• 2017 FPN --Feature Pyramid Networks for Object Detection

• 2017 DSOD --DSOD: Learning Deeply Supervised Object Detectors from Scratch

• 2017 Focal Loss --Focal Loss for Dense Object Detection

• 2017 Mask R-CNN --Mask R-CNN

• 2017 Light-Head R-CNN_In Defense of Two-Stage Object Detector

• TBA

---

2.Dataset&Benchmarks

• Pascal VOC
• ImageNet
• MSCOCO
• KITTI
• CityScapes
• TT100K
• TBA

---

3.Great open code

• Faster-rcnn MatCAFFE,PyCAFFE,Tensorflow,PyTorch,mx-Net

• Deformable-NN Det mx-Net,PyTorch

• R-FCN MatCAFFE,PyCAFFE,PyTorch,CAFFE-MultiGPU

• SSD PyTorch,mx-net,SSD512-pytorch

• YOLO tensorflow,PyTorch,CAFFE,CAFFE

• YOLO2 PyTorch,PyTorch

• DSOD PyCAFFE

• RetinaNet pytorch

• Vehicle Detection & Tracking 1, 2, [3]

• Vehicle Detection and Tracking using HOG, CNN 1

• SSD for KITTI 1

• SqueezeDet on KITTI 1

• Faster-rcnn for KITTI 1

• 3D CNN for KITTI 1

• lane detection based on KITTI model 1

• eval KITTI results 1

• DIDI detection competition 1 2

• Udacity Self-driving Car 1 2 YOLO traffic light detection final project

• RFBNet

• MV3D tensorflow

• Depth and Ego-motion prediction pytorch

• Deeply fuse Net pytorch

---

4.Extra Reading Notes

• Detection Summary Until 17.1 csdn

• Relation Networks for Object Detection zhihu

• SSD details zhihu

• Multi-view 3D detection csdn csdn

• semi-supervised learning on Image zhihu

• CycleGAN,DualGAN,DiscoGAN zhihu

• 人脸遮挡问题的处理 旷视,wechat

---

5.Very Recent Works

a. For 2D object detection

• Feature Selective Networks for Object Detection

• MegDet: A Large Mini-Batch Object Detector

• Deep Image Prior

• An Analysis of Scale Invariance in Object Detection – SNIP

• Receptive Field Block Net for Accurate and Fast Object Detection

• Single-Shot Refinement Neural Network for Object Detection

• Contextual Object Detection with a Few Relevant Neighbors

• Dynamic Zoom-in Network for Fast Object Detection in Large Images

• A Taught-Obesrve-Ask (TOA) Method for Object Detection with Critical Supervision

• Cascade Region Proposal and Global Context for Deep Object Detection

• Relation networks

• Single Shot Text Detector with Regional Attention

• DSSD

• RRC

• GRP-DSOD

• New Salient Object Detection by Lossless Feature Reflection

• New Tiny SSD: A Tiny Single-shot Detection Deep Convolutional Neural Network for Real-time Embedded Object Detection

b. For 3D object detection

• House 3D, a dataset
• Frustum PointNets for 3D Object Detection from RGB-D Data
• VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection
• Single Multi-feature detector for Amodal 3D Object Detection in RGB-D Images

c. For car detection especially

• PointFusion: Deep Sensor Fusion for 3D Bounding Box Estimation
• Joint 3D Proposal Generation and Object Detection from View Aggregation
• Fusing Bird View LIDAR Point Cloud and Front View Camera Image for Deep Object Detection
• A Joint 3D-2D based Method for Free Space Detection on Roads
• PointCNN

d. Other useful tools

• Non-Local