Skip to content

Latest commit

 

History

History
123 lines (75 loc) · 10.5 KB

Paper Survey on Object Detection.md

File metadata and controls

123 lines (75 loc) · 10.5 KB

Bibliography on Object Detection

Literature review of 2D Object Detection

Year Scene Sensor Method Features Link
Jiao et al (Review) 2019 A Survey of Deep Learning-based Object Detection [a1]
He et al (Mask R-CNN) 2017 General RGB Mask R-CNN Two-stage, accurate, generate pixel-wise masks [a2]
Redmon et al (YOLOv2) 2017 General RGB YOLOv2 One-stage, efficient, predict object-wise bounding boxes [a3]

Literature review of 3D object detection with geometry methods.

Scene Sensor Method Model Features Link
parkhiya2018constructing indoor RGB geometry points model matching [b01]
grinvald2019volumetric indoor RGB-D geometry points mask segmentation [b02]
rubino20173d indoor RGB geometry quadric 2D ellipse, multi-view optimization [b03]
nicholson2018quadricslam indoor RGB geometry quadric 2D box to 3D enveloping planes [b04]
liao2020object indoor RGB-D geometry quadric symmetric planes [b05]
gupta2010estimating indoor RGB geometry cuboid line segmentation [b06]
xiao2012localizing indoor RGB geometry cuboid corner detection [b07]
yang2019cubeslam indoor RGB geometry cuboid vanishing lines [b08]
jiang2013linear indoor RGB-D geometry cuboid minimal cuboid [b09]
mishima2019incremental indoor RGB-D geometry cuboid three perpendicular planes [b10]
lin2021topology indoor RGB-D geometry cuboid projecting onto ground [b11]

Literature review of 3D object detection with deep learning methods.

Scene Sensor Method Model Features Link
dwibedi2016deep indoor RGB network cuboid corner regression [c01]
huang2019perspectivenet indoor RGB network cuboid perspective points [c02]
huang2018cooperative indoor RGB network cuboid size, rotation, distance regression, cooperative training [c03]
nie2020total3dunderstanding indoor RGB network cuboid relationship with surrounding [c04]
qi2017pointnet indoor RGB-D network points interesting points selection [c05]
qi2017pointnet++ indoor RGB-D network points geometric structure with neighbourhoods [c06]
qi2019deep indoor RGB-D network points object center voting and geometry constraints [c07]
qi2020imvotenet indoor RGB-D network points 2D-3D consistency, voting schedule [c08]

Others

Scene Sensor Method Model Features Link
lee2009geometric indoor RGB geometry cuboid vanishing line, Hypotheses, 2d to 3d [b12]
hedau2010thinking indoor RGB geometry cuboid GC+HoG, Vanishing line, feature, cuboid model [b13]
lee2010estimating indoor RGB geometry cuboid OM. vanishing lines, orientation map, volumetric constraints [b14]
schwing2013box indoor RGB geometry cuboid GC+OM + DM, Vanishing lines, orientation map, geometric context, branch and bound [b15]
pillai2015monocular indoor RGB geometry cuboid Feature points (SIFT), multi-view, Object Evidence Aggregation [b16]
Li2020view indoor RGB geometry cuboid Feature points, Matching, inlier, line [b17]
wu2020eao indoor RGB geometry cuboid Feature points, Non-parametric test, single-sample test, Ensemble Data Association [b18]
lin2013holistic indoor RGBD geometry cuboid Cluster, Point cloud, multi-view optimization, surface plane estimation [b19]

Reference

[a1] A Survey of Deep Learning-based Object Detection

[a2] Mask R-CNN

[a3] YOLOv2

[b01] Constructing category-specific models for monocular object-slam

[b02] Volumetric instance-aware semantic mapping and 3D object discovery

[b03] 3d object localisation from multi-view image detections

[b04] Quadricslam: Dual quadrics from object detections as landmarks in object-oriented slam

[b05] Object-oriented slam using quadrics and symmetry properties for indoor environments

[b06] Estimating spatial layout of rooms using volumetric reasoning about objects and surfaces

[b07] Localizing 3D cuboids in single-view images

[b08] Cubeslam: Monocular 3-d object slam

[b09] A linear approach to matching cuboids in RGBD images

[b10] Incremental 3d cuboid modeling with drift compensation

[b11] Topology aware object-level semantic mapping towards more robust loop closure

[b12] Geometric reasoning for single image structure recovery

[b13] Thinking inside the box: Using appearance models and context based on room geometry

[b14] Estimating spatial layout of rooms using volumetric reasoning about objects and surfaces

[b15] Box in the box: Joint 3d layout and object reasoning from single images

[b16] Monocular slam supported object recognition

[b17] View-invariant loop closure with oriented semantic landmarks

[b18] Eao-slam: Monocular semi-dense object slam based on ensemble data association

[b19] Holistic scene understanding for 3d object detection with rgbd cameras

[c01] Deep cuboid detection: Beyond 2d bounding boxes

[c02] Perspectivenet: 3d object detection from a single rgb image via perspective points

[c03] Cooperative holistic scene understanding: Unifying 3d object, layout, and camera pose estimation

[c04] Total3dunderstanding: Joint layout, object pose and mesh reconstruction for indoor scenes from a single image

[c05] Pointnet: Deep learning on point sets for 3d classification and segmentation

[c06] Pointnet++: Deep hierarchical feature learning on point sets in a metric space

[c07] Deep hough voting for 3d object detection in point clouds

[c08] Imvotenet: Boosting 3d object detection in point clouds with image votes

[c09] MobilePose: Real-Time Pose Estimation for Unseen Objects with Weak Shape Supervision