2020/10/10に開催された第4回全日本コンピュータビジョン勉強会「人に関する認識・理解論文読み会」発表資料です。
以下の2本を読みました
Harmonious Attention Network for Person Re-identification. (CVPR2018)
Weekly Supervised Person Re-Identification (CVPR2019)
2018/10/20コンピュータビジョン勉強会@関東「ECCV読み会2018」発表資料
Yew, Z. J., & Lee, G. H. (2018). 3DFeat-Net: Weakly Supervised Local 3D Features for Point Cloud Registration. European Conference on Computer Vision.
cvpaper.challenge2019のMeta Study Groupでの発表スライド
点群深層学習についてのサーベイ ( https://www.slideshare.net/naoyachiba18/ss-120302579 )を経た上でのMeta Study
This document summarizes a paper titled "DeepI2P: Image-to-Point Cloud Registration via Deep Classification". The paper proposes a method for estimating the camera pose within a point cloud map using a deep learning model. The model first classifies whether points in the point cloud fall within the camera's frustum or image grid. It then performs pose optimization to estimate the camera pose by minimizing the projection error of inlier points onto the image. The method achieves more accurate camera pose estimation compared to existing techniques based on feature matching or depth estimation. It provides a new approach for camera localization using point cloud maps without requiring cross-modal feature learning.
9. 3D Shapeの表現
9
Figure from “Occupancy Networks: Learning 3D Reconstruction in Function Space”
Voxel Point Cloud Mesh
+Simple
-Cubic Memory
-Manhattan world
+Fast and Easy
-No connectivity
-Lossy Postprocessing
+Natural
-Require Template
(topology)
-Self-intersections
10. 3D Shapeの表現
10
Voxel Point Cloud Mesh Implicit Function
+Infinite Resolution
+Arbitrary Topologies
+Watertight Meshes
Figure from “Occupancy Networks: Learning 3D Reconstruction in Function Space”
+Simple
-Cubic Memory
-Manhattan world
+Fast and Easy
-No connectivity
-Lossy Postprocessing
+Natural
-Require Template
(topology)
-Self-intersections
11. 3D Shapeの表現
11
陰関数(Implicit Function)をDeep Learningで表現
(いずれもCVPR2019)
IM-NET
Learning Implicit Fields for Generative Shape Modeling
OccNET
Occupancy Networks: Learning 3D Reconstruction in
Function Space
DeepSDF
DeepSDF: Learning Continuous Signed Distance
Functions for Shape Representation
https://www.slideshare.net/takmin/20190706cvpr20193dshaperepresentation-153989245
55. Auto-encoding 3D Shapes
55
パーツへ分割する既存研究と比較
Volumetric Primitives (VP)
Tulsiani, S., Su, H., Guibas, L. J., Efros,A.A., & Malik, J. (2017). Learning
shape abstractions by assembling volumetric primitives. In Conference on
ComputerVision and Pattern Recognition.
3D ShapeをPrimitive Shapeの集合で表現
Super Quadrics (SQ)
Paschalidou, D., Ulusoy,A. O., & Geiger,A. (2019). Superquadrics revisited:
Learning 3D shape parsing beyond cuboids. IEEE Conference on Computer
Vision and Pattern Recognition, 2019-June, 10336–10345.
3D Shapeを超楕円体 (Super Quadrics)の集合で表現
Branched Auto Encoders (BAE)
Chen, Z.,Yin, K., Fisher, M., Chaudhuri, S., & Zhang, H. (2019). BAE-NET :
Branched Autoencoder for Shape Co-Segmentation. In International
Conference on ComputerVision.
3D Shapeを陰関数で表現したパーツの集合で表現
61. Single View Reconstruction (SVR)
61
以下の手法と比較
Atlasnet
Groueix,T., Fisher, M., Kim,V. G., Russell, B. C., & Aubry, M. (2018).A
Papier-Mache Approach to Learning 3D Surface Generation. In
Conference on ComputerVision and Pattern Recognition.
OccNet
Mescheder, L., Oechsle, M., Niemeyer, M., Nowozin, S., & Geiger,A.
(2019). Occupancy Networks: Learning 3D Reconstruction in
Function Space. Conference on ComputerVision and Pattern Recognition.
IM-NET
Chen, Z. (2019). Learning Implicit Fields for Generative Shape
Modeling. Conference on ComputerVision and Pattern Recognition.