Footer

Your go-to destination for discovering trending open-source projects and uncovering the insights that matter

Legal

  • Terms and conditions
  • Privacy policy

More

  • Pricing
  • Advertise
  • onedump
  • gosnakego
  • Contact us

Social

  • X (Twitter)
  • GitHub
Trendshift owl
What they build speaks louder than words
© 2026 Trendshift, created by Julian Li with ❤️
SignalAdvertise
Trendshift
SignalAdvertise
Sign in
Trendshift owl

Discover trends that matter

DailyWeeklyMonthlyYearly
Live mentionsTopics
RepositoriesDevelopers
InsightsStats
AxiomCode
AxiomCode
Bifrost
Bifrost
Busbar
Busbar

amap-cvlab/ABot-Recon

Streaming 3D reconstruction from only video input: Revisiting Local Context for Long-Horizon Streaming 3D Reconstruction

Visit GitHub
Python
1.4k
78
3 contributors
Apache License 2.0
website
FeaturedAdvertise to 180K+ monthly visitors
AxiomCode

Give Al Agents a Deeper Understanding of Your Codebase with a Code Graph

Bifrost

Enterprise AI Gateway Built For Scale: Route, Govern, and Secure All Your AI Traffic

Busbar

Control what AI can do before it acts

Momentum

How fast amap-cvlab/ABot-Recon is gaining stars, and where it lands on Trendshift and GitHub Trending

Activity and trending

Stars, forks and merged pull requests each day or month, lined up with the repository's rank on Trendshift Trending

Aug 27 – Oct 19, 2026

Stars +1,384 in 54 days

050100150
101

Forks +78 in 54 days

0510
9

Merged PRs 3 in 54 days

012
2

Trendshift Trendingwhere the repository ranked each day it made the top 25

Not in the Trendshift top 25 in this period

Aug 27Sep 3Sep 10Sep 17Sep 24Oct 1Oct 8Oct 15
The left and right arrow keys step from day to day; Home and End go to the first and the last

Stars, forks, merged pull requests and Trendshift Trending rank, by day
DayStars gainedForks gainedMerged PRs
Thu, Aug 27, 2026+0+00
Fri, Aug 28, 2026+57+40
Sat, Aug 29, 2026+82+90
Sun, Aug 30, 2026+63+50
Mon, Aug 31, 2026+101+40
Tue, Sep 1, 2026+62+10
Wed, Sep 2, 2026+39+60
Thu, Sep 3, 2026+61+50
Fri, Sep 4, 2026+28+20
Sat, Sep 5, 2026+27+40
Sun, Sep 6, 2026+23+20
Mon, Sep 7, 2026+19+10
Tue, Sep 8, 2026+26+00
Wed, Sep 9, 2026+25+10
Thu, Sep 10, 2026+22+10
Fri, Sep 11, 2026+46+30
Sat, Sep 12, 2026+52+30
Sun, Sep 13, 2026+26+10
Mon, Sep 14, 2026+29+10
Tue, Sep 15, 2026+15+00
Wed, Sep 16, 2026+18+30
Thu, Sep 17, 2026+34+20
Fri, Sep 18, 2026+27+20
Sat, Sep 19, 2026+16+00
Sun, Sep 20, 2026+21+00
Mon, Sep 21, 2026+22+20
Tue, Sep 22, 2026+32+00
Wed, Sep 23, 2026+22+10
Thu, Sep 24, 2026+19+10
Fri, Sep 25, 2026+16+00
Sat, Sep 26, 2026+26+20
Sun, Sep 27, 2026+13+00
Mon, Sep 28, 2026+14+10
Tue, Sep 29, 2026+16+10
Wed, Sep 30, 2026+12+10
Thu, Oct 1, 2026+17+00
Fri, Oct 2, 2026+18+10
Sat, Oct 3, 2026+18+10
Sun, Oct 4, 2026+17+00
Mon, Oct 5, 2026+14+00
Tue, Oct 6, 2026+17+00
Wed, Oct 7, 2026+14+00
Thu, Oct 8, 2026+21+00
Fri, Oct 9, 2026+13+00
Sat, Oct 10, 2026+15+10
Sun, Oct 11, 2026+48+30
Mon, Oct 12, 2026+31+10
Tue, Oct 13, 2026+5+00
Wed, Oct 14, 2026+7+11
Thu, Oct 15, 2026+5+00
Fri, Oct 16, 2026+2+02
Sat, Oct 17, 2026+4+00
Sun, Oct 18, 2026+2+10
Mon, Oct 19, 2026+5+00
Total+1,384+783
  • 1#1 on the list
  • 7Rank that day
  • Not trending

Social mentions

Recent discussions about this repository across the web
  1. Sep 161 post
    • ABot-Recon 是高德 CV Lab 开源的实时流式 3D 重建模,只靠最近 12 帧的 KV 缓存上下文窗口,配合上逐帧估计相机坐标系点图和上一帧的位姿串成轨迹,再把局部点图沿轨迹维护在同一坐标系就能持续生成可更新的三维点云地图。 性能方面在 H100 GPU 上 504x280 的分辨率跑 KITTI-02 能到 24.45 FPS。爱芯元智已经把 ABot-Recon 移植到了…

      Read full on X, posted by @hongseyanqiu

      @hongseyanqiu · x.com

  2. Sep 111 post
    • Alibaba just open-sourced an AI that turns any video into a 3D scene, no depth sensor needed. It's called ABot-Recon, from Amap CV Lab (Alibaba). Most streaming reconstruction systems keep growing…

      Read full on X, posted by @simplifyinAI

      @simplifyinAI · x.com

  3. 7 days earlier
  4. Sep 41 post
    • 长视频3D重建一定要不断累积长期记忆吗?ABot-Recon换了个思路:只保留固定12帧局部上下文。 它缓存前11帧的KV特征,逐帧预测点图与相邻位姿,再连续组合成全局轨迹和点云。这样内存和单帧计算量不随视频变长;重访区域还能选用回环优化。关键启发是:长时序任务未必先堆全局状态,也可以把局部预测做稳,再显式组合。

      Read full on X, posted by @vintcessun

      @vintcessun · x.com

  5. Aug 311 post
    • AMAP CV Labが、長時間動画をストリーミング3D再構築する「ABot-Recon」を公開。 参照するのは、直前11フレーム+現在の1フレームだけ。局所点群と隣接フレーム間の姿勢を順次統合し、万フレーム規模の軌跡と3D空間を復元します。 KITTI-02では24.45 FPS、6.71 GiB※。…

      Read full on X, posted by @NarenBao

      @NarenBao · x.com

  6. Aug 301 post
    • Engineered specifically for real-world edge deployment and live robotic perception: • Low Memory Footprint: During the KITTI-02 benchmark tests, the entire reconstruction pipeline required only 6.71…

      Read full on X, posted by @TheTechDiggest

      @TheTechDiggest · x.com

  7. Aug 292 posts
    • What If You Could Reconstruct a Huge 3D World Using Just 12 Frames of Context? ABot-Recon from AMAP CV Lab takes a different approach to long-horizon 3D reconstruction. Instead of storing an…

      Read full on X, posted by @Alacritic_Super

      @Alacritic_Super · x.com

    • 高德开源 ABot-Recon,一个能把长视频实时重建成 3D 世界的模型。 只保留 12 帧局部上下文,就能持续恢复相机轨迹和全局点云,不用把整段视频全塞进显存。 KITTI-02 实测 24.45 FPS、显存 6.71GB。 自动驾驶、机器人、空间智能这类场景,都很有潜力。 代码和模型权重已开源。

      Read full on X, posted by @IndieDevHailey

      @IndieDevHailey · x.com

  8. Aug 281 post
    • AMAP CV Lab introduces ABot-Recon, turning long video streams into camera paths and 3D point clouds in real time. 🤖 🏆 4.35 m ATE and 91.81% F1 on Oxford Spires, without loop closure. ⚡ 24.45 FPS…

      Read full on X, posted by @ModelScope2022

      @ModelScope2022 · x.com

You might also like

Repositories close to amap-cvlab/ABot-Recon
FeaturedAdvertise to 180K+ monthly visitors
AxiomCode

Give Al Agents a Deeper Understanding of Your Codebase with a Code Graph

Bifrost

Enterprise AI Gateway Built For Scale: Route, Govern, and Secure All Your AI Traffic

Busbar

Control what AI can do before it acts

3DAgentWorld/HorizonStream

Python

HorizonStream Official Implementation

25419

Robbyant/lingbot-map

Python

[ECCV 2026 Best Paper Award Candidate] LingBot-Map: Geometric Context Transformer for Streaming 3D Reconstruction

18.1k2.2k

KevinXu02/R3

Python

R^3: 3D Reconstruction via Relative Regression.

34121

Junyi42/LoGeR

Python

Reimplementation of LoGeR: Long-Context Geometric Reconstruction with Hybrid Memory

62649

Haian-Jin/ZipMap

Python

[CVPR 2026] ZipMap: Linear-Time Stateful 3D Reconstruction via Test-Time Training

51119

Inception3D/TTT3R

Python

[ICLR 2026] A simple state update rule to enhance length generalization for CUT3R

73536

zju3dv/Scal3R

Python

[CVPR 2026 (Highlight)] Scal3R: Scalable Test-Time Training for Large-Scale 3D Reconstruction

54539

cwchenwang/tttLRM

Python

[CVPR 2026 Highlight] tttLRM: Test-Time Training for Long Context and Autoregressive 3D Reconstruction

46034

HengyiWang/spann3r

Python

[3DV'25 Award Candidate] 3D Reconstruction with Spatial Memory

1.1k53