2023年7月のポストまとめ

📊 概要

20件のポスト（リプライ除く）を4カテゴリに分類しました。

カテゴリ	件数	割合
🏗️ 3D再構成・SLAM	16	██████████ 80%
🚗 自動運転	2	█ 10%
📄 論文紹介	1	█ 5%
💬 その他	1	█ 5%

🏆 人気トップ3

🥇 1位

RT	35
Like	124
Views	11000

This paper revisits rotation averaging in Structure-from-Motion, proposing to model noise by propagating uncertainty from point correspondences, and integrates MAGSAC loss, improving accuracy.[CVPR 2023] GlobalSfMpy

🔗 ポストを見る

---

🥈 2位

RT	29
Like	94
Views	12000

MonoLaM is a real-time lane mapping algorithm using a monocular camera. It combines 3D detection, lane association, Catmull-Rom Spline parameterization, and optimization to accurately map lanes.(IROS 2023)

🔗 ポストを見る

---

🥉 3位

RT	17
Like	85
Views	7500

An innovative, unsupervised approach improves 3D shape matching by obtaining precise point-wise maps from deep functional maps, effectively handling complex and noisy shapes Unsupervised Learning of Robust Spectral Shape Matching (SIGGRAPH/TOG 2023)

🔗 ポストを見る

---

📂 カテゴリ別ハイライト

🏗️ 3D再構成・SLAM

16件のポスト

MonoLaM is a real-time lane mapping algorithm using a monocular camera. It combines 3D detection, lane association, Catmull-Rom Spline parameterizatio... (♥ 94)

An innovative, unsupervised approach improves 3D shape matching by obtaining precise point-wise maps from deep functional maps, effectively handling c... (♥ 85)

NeRF-Det applies NeRF for 3D detection, using geometry priors for generalizability. It connects detection and NeRF, providing geometry-aware represent... (♥ 62)

• Lab4D is a 4D reconstruction tool using monocular videos, incorporating neural implicit representations, various deformation fields, and several prior… (♥ 62)

A novel diffusion method for 3D generative modeling is introduced, addressing data scarcity and computational complexity issues. This approach, traine... (♥ 60)

🚗 自動運転

2件のポスト

EigenTrajectory (ET) improves trajectory forecasting by transforming pedestrian history into a compact ET space. Using low-rank approximation and traj... (♥ 43)

The study explores uncertainty-aware lidar place recognition, introducing an evaluation protocol and benchmark. It tests five techniques and three dat... (♥ 19)

📄 論文紹介

1件のポスト

This paper revisits rotation averaging in Structure-from-Motion, proposing to model noise by propagating uncertainty from point correspondences, and i... (♥ 124)

💬 その他

1件のポスト

This article introduces visual-inertial navigation (VIN), detailing state estimation, system models, methodologies, calibration, evaluation, and recen... (♥ 65)

📋 全ポスト一覧

全20件を表示

| # | ポスト | RT | ♥ | 日付 | |--:|:-------|---:|--:|:-----| | 1 | [This paper revisits rotation averaging in Structure-from-Motion, proposing to mo...](https://x.com/rsasaki0109/status/1685431312193343488) | 35 | 124 | 2023-07-29 | | 2 | [MonoLaM is a real-time lane mapping algorithm using a monocular camera. It combi...](https://x.com/rsasaki0109/status/1682874286221201408) | 29 | 94 | 2023-07-22 | | 3 | [An innovative, unsupervised approach improves 3D shape matching by obtaining pre...](https://x.com/rsasaki0109/status/1684765356496465920) | 17 | 85 | 2023-07-28 | | 4 | [This article introduces visual-inertial navigation (VIN), detailing state estima...](https://x.com/rsasaki0109/status/1685432111116034048) | 12 | 65 | 2023-07-29 | | 5 | [NeRF-Det applies NeRF for 3D detection, using geometry priors for generalizabili...](https://x.com/rsasaki0109/status/1685039355851898881) | 14 | 62 | 2023-07-28 | | 6 | [Lab4D is a 4D reconstruction tool using monocular videos, incorporating neural i...](https://x.com/rsasaki0109/status/1684277582781612032) | 17 | 62 | 2023-07-26 | | 7 | [A novel diffusion method for 3D generative modeling is introduced, addressing da...](https://x.com/rsasaki0109/status/1685050904297512960) | 13 | 60 | 2023-07-28 | | 8 | [RoITr, a novel Rotation-Invariant Transformer, improves point cloud matching by ...](https://x.com/rsasaki0109/status/1685951977925316608) | 16 | 54 | 2023-07-31 | | 9 | [Tri-Mip encoding factorizes pre-filtered 3D feature spaces into three orthogonal...](https://x.com/rsasaki0109/status/1685076993011068928) | 10 | 47 | 2023-07-28 | | 10 | [Strivec is a 3D representation system using radiance fields and tensor decomposi...](https://x.com/rsasaki0109/status/1685037628318703616) | 12 | 46 | 2023-07-28 | | 11 | [EigenTrajectory (ET) improves trajectory forecasting by transforming pedestrian ...](https://x.com/rsasaki0109/status/1682628928618500096) | 7 | 43 | 2023-07-22 | | 12 | [This work introduces two covariance methods: isotropic, based on predicted score...](https://x.com/rsasaki0109/status/1683316181728649216) | 14 | 40 | 2023-07-24 | | 13 | [WHU-Railway3D: A Diverse Dataset and Benchmark for Railway Point Cloud Semantic ...](https://x.com/rsasaki0109/status/1683259842692276225) | 9 | 37 | 2023-07-23 | | 14 | [(TPAMI2023) Unsupervised Point Cloud Representation Learning with Deep Neural Ne...](https://x.com/rsasaki0109/status/1683313459679883265) | 11 | 33 | 2023-07-24 | | 15 | [This paper presents a graph-theoretic approach for global point cloud registrati...](https://x.com/rsasaki0109/status/1685084412793556992) | 4 | 32 | 2023-07-29 | | 16 | [This study introduces a super-fast LiDAR global localization method, achieving l...](https://x.com/rsasaki0109/status/1685077856551104512) | 7 | 26 | 2023-07-29 | | 17 | [This study introduces a self-supervised learning approach, merging mesh-based fu...](https://x.com/rsasaki0109/status/1685048670490959872) | 8 | 19 | 2023-07-28 | | 18 | [The study explores uncertainty-aware lidar place recognition, introducing an eva...](https://x.com/rsasaki0109/status/1685042289482330113) | 7 | 19 | 2023-07-28 | | 19 | [This study introduces a unique online multi-resolution mapping system, harnessin...](https://x.com/rsasaki0109/status/1685036480975233026) | 3 | 15 | 2023-07-28 | | 20 | [The paper introduces an Occupied Place Description descriptor for SLAM, utilizin...](https://x.com/rsasaki0109/status/1685081385449693184) | 3 | 10 | 2023-07-29 |