Adam W. Harley

272 posts

Adam W. Harley banner
Adam W. Harley

Adam W. Harley

@AdamWHarley

Research Scientist at Meta. I did a postdoc at Stanford and PhD at CMU. I work on computer vision and machine learning.

Redmond, WA Katılım Ağustos 2020
128 Takip Edilen2.3K Takipçiler
Adam W. Harley retweetledi
Xindi Wu
Xindi Wu@cindy_x_wu·
New #NVIDIA Paper We introduce Motive, a motion-centric, gradient-based data attribution method that traces which training videos help or hurt video generation. By isolating temporal dynamics from static appearance, Motive identifies which training videos shape motion in video generation. 🔗 research.nvidia.com/labs/sil/proje… 1/10
English
11
112
541
72.8K
Adam W. Harley retweetledi
Yawar Siddiqui
Yawar Siddiqui@yawarnihal·
Introducing ShapeR, a method for robust conditional 3D shape generation from casually captured sequences. ShapeR leverages a rectified flow transformer conditioned on per-object multimodal data to turn casual image sequences into full metric scene reconstructions. Project Page: facebookresearch.github.io/ShapeR Paper: arxiv.org/abs/2601.11514 Links to code and huggingface below ⬇️
English
17
144
1K
61.2K
Adam W. Harley retweetledi
Dima Damen
Dima Damen@dimadamen·
📢 New Paper PointSt3R: Point Tracking through 3D Grounded Correspondence Can point tracking be re-formulated as pairwise frame correspondence solely? We fine-tuning MASt3R with dynamic correspondences and a visibility loss and achieve competitive point tracking results 1/3
English
3
19
124
10.4K
Yiping Lu
Yiping Lu@2prime_PKU·
Anyone knows adam?
Yiping Lu tweet media
English
265
445
4.8K
633.5K
Adam W. Harley
Adam W. Harley@AdamWHarley·
Alternate sampling rates.
English
0
0
8
713
Adam W. Harley
Adam W. Harley@AdamWHarley·
The visualizer in the AllTracker repo is now GPU-based (much faster), with new options to improve clarity. Very satisfying to see 3D shape "pop out" from the tracking. github.com/aharley/alltra…
English
3
41
327
20.4K
Adam W. Harley
Adam W. Harley@AdamWHarley·
@cam_sentinel @_akhaliq @Gradio Real-time version of this should be doable very soon. On-device is possible, but depends on the device of course. The model shown here is only 16M parameters (66mb), but needs a GPU.
English
0
0
0
57
Cameron Sentinel
Cameron Sentinel@cam_sentinel·
@AdamWHarley @_akhaliq @Gradio Impressive advance-how do you see point tracking optimizing for real-time, on-device inference, especially where cloud connectivity is intermittent or data locality is critical?
English
1
0
1
93
Adam W. Harley
Adam W. Harley@AdamWHarley·
@georgtrof @Gradio Well, it's great failure case for me to stare at, so thanks for sharing! I think point tracking should eventually be the backbone of pose estimation.
English
1
0
0
54
Adam W. Harley
Adam W. Harley@AdamWHarley·
@georgtrof @Gradio I was thinking static ROI yes (e.g., just crop the video with ffmpeg). And yes, blur and self-occlusion make it harder for the model.
English
1
0
0
104
George Trofimov 🚀
George Trofimov 🚀@georgtrof·
@AdamWHarley @Gradio The camera comes with a fixed focus so cannot move too close. Maybe too many issues here - blurry, dynamic plus self occlusions. Or maybe I shouldn't be tracking the object and just use a static roi?
English
1
0
0
83
Adam W. Harley
Adam W. Harley@AdamWHarley·
@mkocab_ @andrew_n_carr @andrew_n_carr I bet you are right that you can classify real/fake from here, but @mkocab_ is right that the visually apparent "cracking" artifact in the background is pretty common overall. This comes from subpixel errors, and maybe subpixel ambiguity in general.
English
0
0
2
64
Muhammed Kocabas
Muhammed Kocabas@mkocab_·
@andrew_n_carr Those carving artifacts don't only happen with generated videos. They occur in real videos too. Here are results from real videos showing the same patterns. This is a failure of the point tracker. Assuming you used AllTracker, @AdamWHarley could probably speak to this better.
English
3
0
3
236
Andrew Carr 🤸
Andrew Carr 🤸@andrew_n_carr·
synthetically generated videos have fewer and fewer visible artifacts. however, they still have substantial noise. one interesting way to detect this noise is to use dense pixel tracking. here are some veo 3 videos with each pixel tracked over time! look at that pattern
English
5
4
48
3.6K
Adam W. Harley
Adam W. Harley@AdamWHarley·
@georgtrof @Gradio Not too bad! A cheap way to improve results here might be to just move the camera closer, or crop and resize the region of interest.
English
1
0
1
371
Adam W. Harley retweetledi
Kwang Moo Yi
Kwang Moo Yi@kwangmoo_yi·
Preprint of (not) today: Harley et al., "AllTracker: Efficient Dense Point Tracking at High Resolution" -- alltracker.github.io Efficient architecture to track all points/pixels in real-time. Matches multiple frames in real time at high resolution via RNN-based refinement.
English
3
23
166
9.8K
Adam W. Harley
Adam W. Harley@AdamWHarley·
Happy to report that AllTracker was accepted to #ICCV2025! The twists and turns and methodical experimentation here took at least 12 months in all. Super hard project, though in retrospect our solution is pretty simple. code: github.com/aharley/alltra… paper: arxiv.org/abs/2506.07310
Adam W. Harley@AdamWHarley

AllTracker: Efficient Dense Point Tracking at High Resolution If you're using any point tracker in any project, this is likely a drop-in upgrade—improving speed, accuracy, and density, all at once.

English
4
10
102
9.6K
Adam W. Harley
Adam W. Harley@AdamWHarley·
AllTracker works by estimating long-range optical flow between a query frame and every other frame in a video, incorporating temporal context in a sliding window. Along with flow, it estimates visibility and confidence, so unreliable tracks can be easily filtered out.
English
1
0
6
1K
Adam W. Harley
Adam W. Harley@AdamWHarley·
AllTracker: Efficient Dense Point Tracking at High Resolution If you're using any point tracker in any project, this is likely a drop-in upgrade—improving speed, accuracy, and density, all at once.
English
2
35
237
29.8K