Hongyang Li (弘洋)

Ph.D. student @SCUT, long-term research intern@IDEA Research, focus on 2D and 3D perception!

[Hongyang Li (弘洋) @ARChaser_]

Nice work from Tingyang🎉. Inspiring idea and impressive performance!

[Chen Wang @chenwangcw]

🚀Introducing Protracker: Inspired by Kalman filter, we tackle point tracking with a robust probabilistic approach.
🌟Our method integrates multiple predictions from both optical flow and semantic correspondences in a unified framework with probabilistic fusion. This ensures to generate smooth and accurate trajectories. ProTracker achieves state-of-the-art performance among self-supervised methods across multiple benchmarks and enhanced robustness in challenging scenarios like occlusion, similar regions, and low-feature areas.
💡Protracker offers a probabilistic framework that combines information of different granularity and semantics, paving the way for new advancements in tracking any point.
🔍#PointTracking #ComputerVision
Project page: https://t.co/9Wsz1QGiX2
Paper: https://t.co/uTOH7Hut9C
Code: https://t.co/9Wsz1QGiX2

[Hongyang Li (弘洋) @ARChaser_]

@torayeff Hi, thank you for your attention. Our paper is currently under review. We will release the code after this stage. 😉

[Hongyang Li (弘洋) @ARChaser_]

🎉 Online Demo
Since TAPTRv3 is an online tracker, recently we have implemented a streaming inference mode, which allows us to process videos of any length on RTX3090!
With the support of @Gradio and @huggingface, we have deployed the demo at https://t.co/XDn4nS7AV4. Try it out!

[JinyuanQu @JinyuanQu322]

💡Introducing TAPTRv3.

[1/3] TAPTRv3 focuses on the robust tracking of any point in long videos. Benefitting from Visibility-aware Long-temporal Attention (VLTA), Context-aware Cross Attention (CCA), and auto-triggered global matching, TAPTRv3 surpasses TAPTRv2 by a large margin

[Hongyang Li (弘洋) @ARChaser_]

The demo has been resumed. 😉

[Hongyang Li (弘洋) @ARChaser_]

The demo is temporarily unavailable due to a server malfunction. We are working urgently to fix the issue. 😥

[Lei Zhang @leizhangcs]

The API for prompt-free object detection is ready, which means you do not need to provide any prompt and DINO-X will automatically recognize, detect, and segment objects in the provided images. Feel free to check out https://t.co/4V7x4tiZax and try this feature.

[Hongyang Li (弘洋) @ARChaser_]

💡Introducing our TAPTRv3

TAPTRv3 focuses on the robust tracking of any point in long videos, surpassing TAPTRv2 by a large margin and achieving SoTA performance. Even when compared with methods trained on internal real-world data, TAPTRv3 is still competitive.

[JinyuanQu @JinyuanQu322]

💡Introducing TAPTRv3.

[1/3] TAPTRv3 focuses on the robust tracking of any point in long videos. Benefitting from Visibility-aware Long-temporal Attention (VLTA), Context-aware Cross Attention (CCA), and auto-triggered global matching, TAPTRv3 surpasses TAPTRv2 by a large margin

[Hongyang Li (弘洋) @ARChaser_]

Welcome to subscribe our official account for more fancy research!😉

[Visincept (Formly IDEA-CVR) @visincept]

With DINO-X, a unified object-centric vision model developed at IDEA-CVR, we can guarantee the best open-world object detection performance to date. #AI #DeepLearning #computerscience #computervision

[Lei Zhang @leizhangcs]

🌟 Introducing DINO-X, our groundbreaking unified vision model at IDEA Research!

Paper: https://t.co/89gUTzuRVV
Blog: https://t.co/zUPG5J4kYx
Playground: https://t.co/KHBfWOyG5P
API: https://t.co/35hbTnR8mK

#AI #ComputerVision #ObjectDetection #IDEAResearch

Highlights:

[AI at Meta @AIatMeta]

New AI research from Meta – CoTracker3 Simpler and Better Point Tracking by Pseudo-Labelling Real Videos.

More details ➡️ https://t.co/b1uoFo7S3g
Demo on @huggingface ➡️ https://t.co/5o5IzC35Nl 

Building on our previous work on CoTracker, this new model demonstrates impressive tracking results where points can be tracked for a long time even when they're occluded or leave the field of view. CoTracker3 achieves state-of-the-art, outperforming all recent point tracking approaches on standard benchmarks — often by a substantial margin.

We've released the research paper, code and a demo on Hugging Face — along with models available under an A-NC license to support further research in this space.

[Hongyang Li (弘洋) @ARChaser_]

@Tianhe_Ren Haha. Thanks, bro 😉

[Hongyang Li (弘洋) @ARChaser_]

#NeurIPS2024
Our TAPTRv2 is accepted by NeurIPS2024. TAPTRv2 is not only simpler and stronger than TAPTR but also unifies the object/point-level tracking framework (for future work). Code will be released soon.

For more details, please refer to our paper. https://t.co/e2HaJjcucR

[Hongyang Li (弘洋) @ARChaser_]

@RHeremans Hi, thank you for your attention, we are going to host our demo on HF, so we disabled the online demo temporally. We are sorry for the inconvenience.

[Hongyang Li (弘洋) @ARChaser_]

We are happy to introduce our new work 「TAPTR: Tracking Any Point with Transformers as Detection」. TAPTR is a simple and strong baseline with sota performance on many datasets while maintaining the advantage of speed.
ProjectPage: https://t.co/o3ha7EWkCV (demos and paper links)

[Hongyang Li (弘洋) @ARChaser_]

@NaveenManwani17 Thank you for your kind sharing!

[Visincept (Formly IDEA-CVR) @visincept]

⭕️CountAnything v1.0.4 is here with some cool updates⭕️
🧸Upload images quicker and enjoy a smoother experience
💨No more hidden SOPs—get to know the APP quickly

more about CountAnything :: https://t.co/1o9fsYDwQP

#CountAnything #Counting #fyp #TwitterWorld

[Visincept (Formly IDEA-CVR) @visincept]

CountAnything promo is coming!
🤨 Developed based on research by @mountch1cken
🫡 Revolutionizing Object Counting Across Generations
🥱Wanna count faster and accurately?

For more information // welcome :: https://t.co/1o9fsYE4Gn

#appsforcounting

[Visincept (Formly IDEA-CVR) @visincept]

⚠️T-rex Label v1.2 Update is here⚠️
✅You can now manually get rid of extra boxes in smart drawing
✅Shortcut keys have been improved
✅Creating categories is faster and easier

more about product? :: https://t.co/968qhdQMpZ

#computersciences #CV #annotation #TREX

[Hongyang Li (弘洋) @ARChaser_]

@JieWang_ZJUI 😂 Thank you Jie, hope you will like this tool. In fact 'pudb' could be more fancy, you can also have a try.

[Hongyang Li (弘洋) @ARChaser_]

Awesome T-Rex, welcome to experience T-Rex.

[Visincept (Formly IDEA-CVR) @visincept]

Don't Miss Out! T-Rex Label in various industrial detection tasks! #computervision #Data #Analysis #GenerativeAI #deeplearning #annotation #fyp #Today #TRY
Try it now: https://t.co/KaOmoUDorK