objectnet3d: a large scale database for 3d object...
TRANSCRIPT
![Page 1: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/1.jpg)
ObjectNet3D: A Large Scale Database for
3D Object Recognition
Yu Xiang, Wonhui Kim, Wei Chen, Jingwei Ji, Christopher Choy,
Hao Su, Roozbeh Mottaghi, Leonidas Guibas and Silvio Savarese
ECCV 2016
1
![Page 2: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/2.jpg)
Recognizing the 3D Properties of Objects
• 3D location, 3D pose, 3D shape, etc.
• Applications
2
Robotics Autonomous
Driving
Augmented
Reality
![Page 3: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/3.jpg)
Our Contribution: ObjectNet3D Database
• A large scale database for 3D object recognition
3
![Page 4: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/4.jpg)
3D Annotation: 2D-3D Alignment
4
![Page 5: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/5.jpg)
3D Annotation: 2D-3D Alignment
5
![Page 6: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/6.jpg)
3D Annotation: 2D-3D Alignment
6
![Page 7: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/7.jpg)
3D Annotation: 2D-3D Alignment
7
![Page 8: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/8.jpg)
3D Annotation: 2D-3D Alignment
8
![Page 9: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/9.jpg)
Comparison with Previous Datasets
9
#category #instance Non-centered
objects
Dense
viewpoint
3D Shape
3D Object [1] 10 100
EPFL Car [2] 1 20 ✓
RGB-D Object [3] 51 300 ✓
PASCAL VOC [4] 20 27,450 ✓
KITTI [5] 3 80,256 ✓ ✓
PASCAL3D+ [6] 12 35,672 ✓ ✓ ✓79
[1] S. Savarese and L. Fei-Fei. 3d generic object categorization, localization and pose estimation. In ICCV, 2007.
[2] M. Ozuysal, V. Lepetit, and P. Fua. Pose estimation for category specific multiview object localization. In CVPR, 2009.[3] K. Lai, L. Bo, X. Ren and D. Fox. A large-scale hierarchical multi-view RGB-D object dataset. In ICRA, 2011.
[4] M. Everingham, L. Van Gool, C. K. I.Williams, J.Winn, and A. Zisserman. The pascal visual object classes (voc) challenge. IJCV, 2010.
[5] A. Geiger, P. Lenz, and R. Urtasun. Are we ready for autonomous driving? the kitti vision benchmark suite. In CVPR, 2012.[6] Y. Xiang, R. Mottaghi and S. Savarese. Beyond PASCAL: A benchmark for 3D object detection in the wild. In WACV, 2014.
![Page 10: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/10.jpg)
Comparison with Previous Datasets
10
#category #instance Non-centered
objects
Dense
viewpoint
3D Shape
3D Object [1] 10 100
EPFL Car [2] 1 20 ✓
RGB-D Object [3] 51 300 ✓
PASCAL VOC [4] 20 27,450 ✓
KITTI [5] 3 80,256 ✓ ✓
PASCAL3D+ [6] 12 35,672 ✓ ✓ ✓79
[1] S. Savarese and L. Fei-Fei. 3d generic object categorization, localization and pose estimation. In ICCV, 2007.
[2] M. Ozuysal, V. Lepetit, and P. Fua. Pose estimation for category specific multiview object localization. In CVPR, 2009.[3] K. Lai, L. Bo, X. Ren and D. Fox. A large-scale hierarchical multi-view RGB-D object dataset. In ICRA, 2011.
[4] M. Everingham, L. Van Gool, C. K. I.Williams, J.Winn, and A. Zisserman. The pascal visual object classes (voc) challenge. IJCV, 2010.
[5] A. Geiger, P. Lenz, and R. Urtasun. Are we ready for autonomous driving? the kitti vision benchmark suite. In CVPR, 2012.[6] Y. Xiang, R. Mottaghi and S. Savarese. Beyond PASCAL: A benchmark for 3D object detection in the wild. In WACV, 2014.
![Page 11: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/11.jpg)
Comparison with Previous Datasets
11
#category #instance Non-centered
objects
Dense
viewpoint
3D Shape
3D Object [1] 10 100
EPFL Car [2] 1 20 ✓
RGB-D Object [3] 51 300 ✓
PASCAL VOC [4] 20 27,450 ✓
KITTI [5] 3 80,256 ✓ ✓
PASCAL3D+ [6] 12 35,672 ✓ ✓ ✓79
[1] S. Savarese and L. Fei-Fei. 3d generic object categorization, localization and pose estimation. In ICCV, 2007.
[2] M. Ozuysal, V. Lepetit, and P. Fua. Pose estimation for category specific multiview object localization. In CVPR, 2009.[3] K. Lai, L. Bo, X. Ren and D. Fox. A large-scale hierarchical multi-view RGB-D object dataset. In ICRA, 2011.
[4] M. Everingham, L. Van Gool, C. K. I.Williams, J.Winn, and A. Zisserman. The pascal visual object classes (voc) challenge. IJCV, 2010.
[5] A. Geiger, P. Lenz, and R. Urtasun. Are we ready for autonomous driving? the kitti vision benchmark suite. In CVPR, 2012.[6] Y. Xiang, R. Mottaghi and S. Savarese. Beyond PASCAL: A benchmark for 3D object detection in the wild. In WACV, 2014.
![Page 12: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/12.jpg)
Comparison with Previous Datasets
12
#category #instance Non-centered
objects
Dense
viewpoint
3D Shape
3D Object [1] 10 100
EPFL Car [2] 1 20 ✓
RGB-D Object [3] 51 300 ✓
PASCAL VOC [4] 20 27,450 ✓
KITTI [5] 3 80,256 ✓ ✓
PASCAL3D+ [6] 12 35,672 ✓ ✓ ✓79
[1] S. Savarese and L. Fei-Fei. 3d generic object categorization, localization and pose estimation. In ICCV, 2007.
[2] M. Ozuysal, V. Lepetit, and P. Fua. Pose estimation for category specific multiview object localization. In CVPR, 2009.[3] K. Lai, L. Bo, X. Ren and D. Fox. A large-scale hierarchical multi-view RGB-D object dataset. In ICRA, 2011.
[4] M. Everingham, L. Van Gool, C. K. I.Williams, J.Winn, and A. Zisserman. The pascal visual object classes (voc) challenge. IJCV, 2010.
[5] A. Geiger, P. Lenz, and R. Urtasun. Are we ready for autonomous driving? the kitti vision benchmark suite. In CVPR, 2012.[6] Y. Xiang, R. Mottaghi and S. Savarese. Beyond PASCAL: A benchmark for 3D object detection in the wild. In WACV, 2014.
![Page 13: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/13.jpg)
Comparison with Previous Datasets
13
#category #instance Non-centered
objects
Dense
viewpoint
3D Shape
3D Object [1] 10 100
EPFL Car [2] 1 20 ✓
RGB-D Object [3] 51 300 ✓
PASCAL VOC [4] 20 27,450 ✓
KITTI [5] 3 80,256 ✓ ✓
PASCAL3D+ [6] 12 35,672 ✓ ✓ ✓79
[1] S. Savarese and L. Fei-Fei. 3d generic object categorization, localization and pose estimation. In ICCV, 2007.
[2] M. Ozuysal, V. Lepetit, and P. Fua. Pose estimation for category specific multiview object localization. In CVPR, 2009.[3] K. Lai, L. Bo, X. Ren and D. Fox. A large-scale hierarchical multi-view RGB-D object dataset. In ICRA, 2011.
[4] M. Everingham, L. Van Gool, C. K. I.Williams, J.Winn, and A. Zisserman. The pascal visual object classes (voc) challenge. IJCV, 2010.
[5] A. Geiger, P. Lenz, and R. Urtasun. Are we ready for autonomous driving? the kitti vision benchmark suite. In CVPR, 2012.[6] Y. Xiang, R. Mottaghi and S. Savarese. Beyond PASCAL: A benchmark for 3D object detection in the wild. In WACV, 2014.
![Page 14: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/14.jpg)
Comparison with Previous Datasets
14
#category #instance Non-centered
objects
Dense
viewpoint
3D Shape
3D Object [1] 10 100
EPFL Car [2] 1 20 ✓
RGB-D Object [3] 51 300 ✓
PASCAL VOC [4] 20 27,450 ✓
KITTI [5] 3 80,256 ✓ ✓
PASCAL3D+ [6] 12 35,672 ✓ ✓ ✓79
[1] S. Savarese and L. Fei-Fei. 3d generic object categorization, localization and pose estimation. In ICCV, 2007.
[2] M. Ozuysal, V. Lepetit, and P. Fua. Pose estimation for category specific multiview object localization. In CVPR, 2009.[3] K. Lai, L. Bo, X. Ren and D. Fox. A large-scale hierarchical multi-view RGB-D object dataset. In ICRA, 2011.
[4] M. Everingham, L. Van Gool, C. K. I.Williams, J.Winn, and A. Zisserman. The pascal visual object classes (voc) challenge. IJCV, 2010.
[5] A. Geiger, P. Lenz, and R. Urtasun. Are we ready for autonomous driving? the kitti vision benchmark suite. In CVPR, 2012.[6] Y. Xiang, R. Mottaghi and S. Savarese. Beyond PASCAL: A benchmark for 3D object detection in the wild. In WACV, 2014.
![Page 15: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/15.jpg)
Comparison with Previous Datasets
15
#category #instance Non-centered
objects
Dense
viewpoint
3D Shape
3D Object [1] 10 100
EPFL Car [2] 1 20 ✓
RGB-D Object [3] 51 300 ✓
PASCAL VOC [4] 20 27,450 ✓
KITTI [5] 3 80,256 ✓ ✓
PASCAL3D+ [6] 12 35,672 ✓ ✓ ✓79
ObjectNet3D (Ours) 100 201,888 ✓ ✓ ✓44,147
[1] S. Savarese and L. Fei-Fei. 3d generic object categorization, localization and pose estimation. In ICCV, 2007.
[2] M. Ozuysal, V. Lepetit, and P. Fua. Pose estimation for category specific multiview object localization. In CVPR, 2009.[3] K. Lai, L. Bo, X. Ren and D. Fox. A large-scale hierarchical multi-view RGB-D object dataset. In ICRA, 2011.
[4] M. Everingham, L. Van Gool, C. K. I.Williams, J.Winn, and A. Zisserman. The pascal visual object classes (voc) challenge. IJCV, 2010.
[5] A. Geiger, P. Lenz, and R. Urtasun. Are we ready for autonomous driving? the kitti vision benchmark suite. In CVPR, 2012.[6] Y. Xiang, R. Mottaghi and S. Savarese. Beyond PASCAL: A benchmark for 3D object detection in the wild. In WACV, 2014.
![Page 16: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/16.jpg)
Database Construction: Object Categories
• 100 rigid object categories
16
Aeroplane
Ashtray
Backpack
Basket
Bed
Bench
Bicycle
Backboard
Boat
Bookshelf
Bottle
Bucket
Bus
Cabinet
Calculator
Camera
Can
Cap
Car
Cellphone
Chair
Clock
Coffee maker
Comb
Computer
Cup
Desk lamp
Dining table
Dishwasher
Door
Eraser
Eyeglasses
Fan
Faucet
Filing cabinet
Fire extinguisher
Fish tank
Flashlight
Fork
Guitar
Hair dryer
Hammer
Headphone
Helmet
Iron
Jar
Kettle
Key
Keyboard
Knife
Laptop
Lighter
Mailbox
Microphone
Microwave
Motorbike
Mouse
Paintbrush
Pan
Pen
Pencil
Piano
Pillow
Plate
Pot
Printer
Racket
Refrigerator
Remote control
Rifle
Road pole
Satellite dish
Scissors
Screwdriver
Shoe
Shovel
Sign
Skate
Skateboard
Slipper
Sofa
Speaker
Spoon
Stapler
Stove
Suitcase
Teapot
Telephone
Toaster
Toilet
Toothbrush
Train
Trash bin
Trophy
Tub
Tvmonitor
Vending machine
Washing machine
Watch
Wheelchair
![Page 17: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/17.jpg)
• 100 rigid object categories
17
Aeroplane
Ashtray
Backpack
Basket
Bed
Bench
Bicycle
Backboard
Boat
Bookshelf
Bottle
Bucket
Bus
Cabinet
Calculator
Camera
Can
Cap
Car
Cellphone
Chair
Clock
Coffee maker
Comb
Computer
Cup
Desk lamp
Dining table
Dishwasher
Door
Eraser
Eyeglasses
Fan
Faucet
Filing cabinet
Fire extinguisher
Fish tank
Flashlight
Fork
Guitar
Hair dryer
Hammer
Headphone
Helmet
Iron
Jar
Kettle
Key
Keyboard
Knife
Laptop
Lighter
Mailbox
Microphone
Microwave
Motorbike
Mouse
Paintbrush
Pan
Pen
Pencil
Piano
Pillow
Plate
Pot
Printer
Racket
Refrigerator
Remote control
Rifle
Road pole
Satellite dish
Scissors
Screwdriver
Shoe
Shovel
Sign
Skate
Skateboard
Slipper
Sofa
Speaker
Spoon
Stapler
Stove
Suitcase
Teapot
Telephone
Toaster
Toilet
Toothbrush
Train
Trash bin
Trophy
Tub
Tvmonitor
Vending machine
Washing machine
Watch
Wheelchair
Vehicles Furniture Container
Tools Electronics Personal items
Database Construction: Object Categories
![Page 18: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/18.jpg)
Database Construction: Images
• 2D images from the ImageNet database [1]
18[1] Russakovsky et al. ImageNet Large Scale Visual Recognition Challenge, IJCV 2015
![Page 19: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/19.jpg)
Database Construction: 3D Shapes
• Trimble 3D Warehouse [1]
• ShapeNet database [2]
193D Shapes from Trimble 3D Warehouse 3D Shapes from ShapeNet
[2] Chang et al. ShapeNet: An Information-Rich 3D Model Repository, arXiv 2015[1] https://3dwarehouse.sketchup.com
![Page 20: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/20.jpg)
Database Construction: Annotation
Demo
20
![Page 21: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/21.jpg)
3D Pose Annotation Examples
21
![Page 22: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/22.jpg)
Viewpoint Distributions
22
aeroplane bed cup
mouseeyeglasses
![Page 23: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/23.jpg)
Database Construction: Image-based
3D Shape Retrieval
23
![Page 24: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/24.jpg)
Database Construction: Image-based
3D Shape Retrieval
24
![Page 25: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/25.jpg)
Database Construction: Image-based
3D Shape Retrieval
25
![Page 26: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/26.jpg)
Database Construction: Image-based
3D Shape Retrieval
26
![Page 27: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/27.jpg)
Database Construction: Image-based
3D Shape Retrieval
27
![Page 28: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/28.jpg)
Database Construction: Image-based
3D Shape Retrieval
28
Test Object
![Page 29: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/29.jpg)
Database Construction: Image-based
3D Shape Retrieval
29H.O. Song, Y. Xiang, S. Jegelka and S. Savarese. Deep Metric Learning via Lifted Structured Feature Embedding. In CVPR, 2016.
Test Object Rank 1 Rank 2 Rank 3
…
…
…
![Page 30: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/30.jpg)
Database Construction: Image-based
3D Shape Retrieval
30H.O. Song, Y. Xiang, S. Jegelka and S. Savarese. Deep Metric Learning via Lifted Structured Feature Embedding. In CVPR, 2016.
Test Object Rank 1 Rank 2 Rank 3
…
…
…
![Page 31: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/31.jpg)
Baseline Experiments
• Object proposal generation
• 2D object detection
• Image-based 3D shape retrieval
• Joint 2D detection and continuous 3D pose estimation
31
![Page 32: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/32.jpg)
Baseline Experiments
• Object proposal generation
• 2D object detection
• Image-based 3D shape retrieval
• Joint 2D detection and continuous 3D pose estimation
32
Selective Search: Uijlings et al., IJCV, 2013.
EdgeBoxes: Zitnick et al., ECCV, 2014.MCG: Arbelaez et al., CVPR, 2014.RPN: Ren et al., NIPS, 2015.
![Page 33: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/33.jpg)
Baseline Experiments
• Object proposal generation
• 2D object detection
• Image-based 3D shape retrieval
• Joint 2D detection and continuous 3D pose estimation
33
Selective Search: Uijlings et al., IJCV, 2013.
EdgeBoxes: Zitnick et al., ECCV, 2014.MCG: Arbelaez et al., CVPR, 2014.RPN: Ren et al., NIPS, 2015.
Fast R-CNN: Girshick R., ICCV, 2015.
![Page 34: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/34.jpg)
Baseline Experiments
• Object proposal generation
• 2D object detection
• Image-based 3D shape retrieval
• Joint 2D detection and continuous 3D pose estimation
34
Selective Search: Uijlings et al., IJCV, 2013.
EdgeBoxes: Zitnick et al., ECCV, 2014.MCG: Arbelaez et al., CVPR, 2014.RPN: Ren et al., NIPS, 2015.
Fast R-CNN: Girshick R., ICCV, 2015.
Deep Lifted Structure: Song et al., CVPR, 2016.
![Page 35: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/35.jpg)
A Network for Object Detection and
Pose estimation
35R. Girshick. Fast R-CNN. ICCV’15.
![Page 36: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/36.jpg)
A Network for Object Detection and
Pose estimation
36R. Girshick. Fast R-CNN. ICCV’15.
![Page 37: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/37.jpg)
ObjectNet3D
37
100 object categories
90,127 images
201,888 objects
44,147 3D shapes
2D-3D alignments
Baseline experiments
on different
recognition tasks
![Page 38: ObjectNet3D: A Large Scale Database for 3D Object Recognitionyuxng.github.io/xiang_eccv16_slides.pdf · ObjectNet3D: A Large Scale Database for 3D Object Recognition Yu Xiang, Wonhui](https://reader036.vdocument.in/reader036/viewer/2022063009/5fc158853c362c4836695bd4/html5/thumbnails/38.jpg)
Thank you!